<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Max Boguslavskiy on Medium]]></title>
        <description><![CDATA[Stories by Max Boguslavskiy on Medium]]></description>
        <link>https://medium.com/@maxbogus?source=rss-63caf087450b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*2lG9_XLfh325blot4UX6VA.jpeg</url>
            <title>Stories by Max Boguslavskiy on Medium</title>
            <link>https://medium.com/@maxbogus?source=rss-63caf087450b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 23:23:52 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@maxbogus/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Innovation in Russia: Can Agile Defeat the System?]]></title>
            <link>https://maxbogus.medium.com/innovation-in-russia-can-agile-defeat-the-system-7e4be706d4fb?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/7e4be706d4fb</guid>
            <category><![CDATA[transformation]]></category>
            <category><![CDATA[russia]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Tue, 25 Nov 2025 20:59:30 GMT</pubDate>
            <atom:updated>2025-11-25T20:59:30.140Z</atom:updated>
            <content:encoded><![CDATA[<p>Innovation in Russia: Can Agile Defeat the System?<br>A Data-Driven Analysis with the &quot;5 Metafactors&quot; Framework</p><p>Extended Version with Actionable Steps</p><p>---</p><p>Let&#39;s be honest: Russia is one of the most challenging countries for Agile implementation.</p><p>And the reason isn&#39;t management. It&#39;s in the cultural code that has been forming for centuries.</p><p>### What the Data Says</p><p>Hofstede Insights gives Russia a score of 95 out of 100 on the Uncertainty Avoidance Index (UAI). This is one of the highest scores in the world. Only Greece (100) and Portugal (99) rank higher.</p><p>For comparison:<br>- USA — 46<br>- China — 30<br>- Singapore — 8</p><p>What does this mean? Russians experience significant discomfort with uncertainty and actively create systems to reduce it: bureaucracy, formal procedures, detailed planning[151][142].</p><p>And here&#39;s the problem: Agile is a methodology that embraces uncertainty as a given. Iterations. Changing requirements. &quot;Fail fast, learn fast.&quot;</p><p>For a culture with UAI 95, this sounds like chaos.</p><p>---</p><p>## THE 5 METAFACTORS: Why Culture Matters More Than Process</p><p>Before we can implement Agile successfully in Russia, we need to understand what actually drives innovation in large organizations. Based on Hofstede, McKinsey, Gallup, and SHRM research, there are 5 metafactors that determine whether an organization can sustain innovation culture[254][255][257][259][260]:</p><p>### Metafactor 1: ORGANIZATIONAL SIZE (Impact: 5/5)<br>- &lt;150 people: 95-100% innovation probability<br>- 150-500 people: 75-90%<br>- 500-1,500 people: 50-75% (Critical Point)<br>- &gt;3,000 people: 0-25%</p><p>Why it matters: At 500+ headcount, you physically cannot maintain a flat hierarchy. You need structure, which naturally creates bureaucracy.</p><p>For Russia specifically: With average company size 2,500+, we&#39;re fighting organizational physics. The solution isn&#39;t to ignore hierarchy—it&#39;s to use it strategically.</p><p>### Metafactor 2: LEADERSHIP STYLE (Impact: 5/5)<br>- eNPS &gt;+30 (Servant leadership): Innovation survives even at 20,000 employees<br>- eNPS 0 to +30 (Matrix style): Conditional—depends on empowerment<br>- eNPS &lt;0 (Command &amp; Control): Innovation dies, becomes bureaucracy</p><p>The Russian Challenge: Our cultural default is command-and-control (Power Distance Index: 93)[151][142]. But eNPS &gt;+30 is achievable—Google and Netflix prove it.</p><p>Data point: Companies with eNPS &gt;+20 maintain 2.5x higher innovation rates than those with eNPS &lt;0[259].</p><p>### Metafactor 3: EMPLOYEE AUTONOMY (Impact: 5/5)<br>- &gt;40% of employees initiate projects: Startup culture maintained<br>- 20-40%: Hybrid—some innovation, mostly execution<br>- &lt;20%: Corporate—people follow instructions</p><p>Why Agile fails in Russia: We tell people &quot;self-organize,&quot; but our culture says &quot;wait for instructions.&quot; 20% UAI breeds compliance, not autonomy.</p><p>Success case: Google&#39;s &quot;20% time&quot; policy directly increased autonomy from 15% to 42% of engineers initiating projects[322][325][331].</p><p>### Metafactor 4: INNOVATION OUTPUT (Impact: 5/5)<br>- &gt;20% revenue from products &lt;5 years old: Innovative company<br>- 10-20%: Transitional<br>- &lt;10%: Conservative</p><p>The Gap in Russia: Concept-2020 target was 40-50% of enterprises innovating. Actual: 21.5%[306]. R&amp;D spending target: 3% GDP. Actual: 1.1%[306].</p><p>This isn&#39;t a capability problem—it&#39;s a system problem. We can innovate, but our structures block it.</p><p>### Metafactor 5: AGILITY (Impact: 4/5)<br>- &lt;2 weeks between releases: Startup agility<br>- 2-4 weeks: Manageable<br>- &gt;4 weeks: Corporate bureaucracy</p><p>The Russian Reality: State-controlled companies release quarterly. Private tech companies (Yandex, Avito) do 1-2 weeks. The difference? Who&#39;s controlling the risk aversion.</p><p>---</p><p>## Historical Perspective: Why This Pattern Keeps Repeating</p><p>Russia&#39;s struggle against &quot;wrong&quot; culture in pursuit of modernization is nothing new.</p><p>Peter the Great (1696-1725) conducted radical reforms: created the Table of Ranks (1722), introduced colleges instead of chancellories, reformed the military along European lines, and founded Saint Petersburg as a &quot;window to Europe&quot;[288][290][301]. He literally forced nobles to shave their beards and wear European clothes, imposing a beard tax on those who refused[288].</p><p>The result? Russia became Europe&#39;s largest iron producer by the end of his reign[290]. But the cultural code remained: hierarchy, centralized decision-making, distrust of change.</p><p>300 years later we see the same pattern: innovations are implemented top-down, through decrees and strategies.</p><p>Here&#39;s the insight: Peter didn&#39;t work *against* culture—he used hierarchy itself as the lever. He didn&#39;t ask nobles to &quot;self-organize.&quot; He commanded. And it worked.</p><p>The lesson for modern Agile in Russia? Stop pretending we&#39;re California startups. Lean into what works culturally.</p><p>---</p><p>## The Paradox: Education Exists, Implementation Doesn&#39;t</p><p>Russia is a country with a high level of education. According to OECD data, the proportion of the population with higher education in Russia exceeds the European average. Russian students are traditionally strong in mathematics and natural sciences.</p><p>Hofstede notes: Russia is a &quot;feminine&quot; culture where education is valued for its own sake, and modesty regarding achievements is considered normal[142][314].</p><p>But here&#39;s the paradox: a high level of education doesn&#39;t convert to innovation.</p><p>Why? Because innovation requires not just knowledge, but tolerance for uncertainty. And that&#39;s precisely what the cultural code blocks.</p><p>The 5 Metafactors Framework explains this gap:<br>- We have Metafactor 2 (Leadership) and Metafactor 3 (Autonomy) misaligned<br>- We optimize for Metafactor 1 (Size) through hierarchy, which kills innovation<br>- Result: Smart people, blocked systems</p><p>---</p><p>## What Actually Works: The Metafactors in Action</p><p>Research on Agile transformation cases in Russian companies shows: successful implementations work not against the 5 metafactors, but through them[322][325][331].</p><p>BioCad (pharma, 2,500 employees) implemented Agile in R&amp;D divisions by[325]:</p><p>1. Using Metafactor 2 (Leadership): Top management support. Not &quot;flat,&quot; but &quot;empowering hierarchy.&quot;<br>2. Increasing Metafactor 3 (Autonomy): Pilot with volunteers (40% initiated projects in pilot)<br>3. Protecting Metafactor 5 (Agility): Daily standups, weekly releases (vs quarterly)<br>4. Managing Metafactor 1 (Size): Split R&amp;D into 3 independent pods (each &lt;200 people)</p><p>Raiffeisenbank Russia successfully conducted digital transformation by[331]:<br>- Measuring eNPS (Metafactor 2) and using it as primary metric<br>- Publishing autonomy metrics (% project initiations)<br>- Decentralizing into smaller teams</p><p>Key insight: In Russia, Agile works when you measure the 5 metafactors, acknowledge constraints, and work within them**—not against them.</p><p>---</p><p>## Stubbornness as an Asset</p><p>Here&#39;s what&#39;s interesting: Hofstede also shows that Russia has a high Long-Term Orientation score (81 out of 100)[145][287]. This means: persistence, pragmatism, ability to adapt traditions to changing conditions.</p><p>The Chinese call Russians a &quot;martial people&quot; — in a positive sense: endurance, willpower, drive to overcome obstacles[326].</p><p>Research on GRIT (perseverance and passion for long-term goals) in Russia shows: Russian students demonstrate high levels of persistence in achieving educational goals[323][337].</p><p>**In other words: if Russians decide to do something — they&#39;ll do it. The question is whether they decided to.</p><p>This is actually our greatest asset for Agile transformation. We don&#39;t need to change culture overnight. We need to commit to a 3-5 year journey and execute with Russian persistence.</p><p>---</p><p>## The Dependency Model: How Metafactors Drive Innovation Probability</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cz_lTQ75EX_4YXpSNGjR7Q.png" /><figcaption>Initial matrix to check company for innovation culture that was reworked into 5 Metafactors Probability Matrix</figcaption></figure><p>Here&#39;s where the 5 Metafactors framework becomes actionable. Each metafactor contributes to overall innovation probability:</p><p>INNOVATION PROBABILITY = <br> (Size_Factor × 0.20) + <br> (Leadership_eNPS × 0.25) + <br> (Autonomy_% × 0.25) + <br> (Innovation_Output × 0.20) + <br> (Agility_Weeks × 0.10)</p><p>For a typical Russian tech company (2,500 people):<br>- Size: 30% (large, constrained)<br>- Leadership eNPS: 35% (if eNPS = +10, vs target +30)<br>- Autonomy: 45% (if 25% initiate projects, vs target 40%)<br>- Innovation Output: 40% (if 12% revenue from new products, vs target 20%)<br>- Agility: 55% (if 3-week releases, vs target &lt;2 weeks)</p><p>Weighted probability: ~40-45% (TRANSITIONAL STATE)</p><p>This matches our earlier assessment. But now we know exactly which metafactor to move first.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LXTrFzOt_-lPlnOJR9QM6g.png" /><figcaption>Startup Spirit Probability Matrix</figcaption></figure><p>## The Russian Advantage: Why We Can Actually Win</p><p>The paradox is also our strength:</p><p>1. High education level = can understand complex methodologies<br>2. High UAI (95) = if we clarify the system, people follow precisely<br>3. High LTO (81) = persistence for 3-5 year transformation<br>4. Hierarchy acceptance = top-down mandates work (unlike Western resistance)<br>5. Engineering excellence = we *can* execute, just need permission and clarity</p><p>Peter the Great didn&#39;t ask nobles to &quot;embrace change.&quot; He commanded it, clarified why, and executed systematically.</p><p>We can do the same with Agile—if we stop pretending we&#39;re Silicon Valley and start acting like Russia.</p><p>---</p><p>## Conclusion</p><p>Russia is not &quot;incapable&quot; of innovation. Russia requires a different approach to its implementation—one that works *with* the 5 Metafactors, not against them:</p><p>1. Use hierarchy as a lever (Metafactor 2: Leadership)<br>2. Protect autonomy through structure (Metafactor 3: Autonomy)<br>3. Divide and empower smaller teams (Metafactor 1: Size)<br>4. Measure everything explicitly (Metafactor 4: Innovation Output)<br>5. Accelerate through small cycles (Metafactor 5: Agility)</p><p>Start this week. Commit for 12 months. Measure monthly.</p><p>Agile can defeat the system. But not through negating the cultural code—**by using it as the engine of change.</p><p>The question isn&#39;t &quot;Will Russia innovate?&quot; The question is: &quot;Are we ready to lead the transformation?&quot;</p><p>---</p><p>**#OrganizationalCulture #AgileTransformation #Russia #5Metafactors #Innovation #DataDriven #Leadership</p><p>---</p><p>## Sources:</p><p>1. Hofstede Insights — Russia Uncertainty Avoidance Index (95/100), Long-Term Orientation (81/100)<br>2. McKinsey — &quot;Organizational Health Index&quot; and cultural dimensions in innovation<br>3. Gallup — eNPS research on employee engagement and innovation correlation<br>4. SHRM — Cost of Turnover and organizational culture studies<br>5. Competing Values Framework (Quinn, 1988) — Applied to Russian context<br>6. Klepach A.N. (2022) — &quot;Russian Science and Technology: Rise or Progressive Lag?&quot; — PMC<br>7. Volkova N.V., Plakhotnik M.S. (2022) — &quot;Overcoming Barriers to Agile at BioCad&quot; — HSE Graduate School<br>8. INSEAD Case — &quot;Agile Transformation of Raiffeisenbank: Culture First&quot;<br>9. ProAgile.ru — &quot;15 real cases of Agile transformation in Russian companies&quot;<br>10. Tovar-García E.D. (2017) — &quot;The impact of perseverance and passion for long term goals&quot; — ScienceDirect</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7e4be706d4fb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Transpiling Typescript into double packages (CommonJS + ESM)]]></title>
            <link>https://levelup.gitconnected.com/transpiling-typescript-into-double-packages-commonjs-esm-d6b62cfc851c?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/d6b62cfc851c</guid>
            <category><![CDATA[typescript]]></category>
            <category><![CDATA[commonjs]]></category>
            <category><![CDATA[es6]]></category>
            <category><![CDATA[nodejs]]></category>
            <category><![CDATA[esm]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Thu, 22 Sep 2022 15:36:28 GMT</pubDate>
            <atom:updated>2022-09-22T15:36:28.949Z</atom:updated>
            <content:encoded><![CDATA[<h3>Mistakes. They are everywhere</h3><p>The reason for writing the article was errors like:</p><blockquote>Error [ERR_REQUIRE_ESM]: require() of ES Module …/index.js from …/file.ts not supported.</blockquote><blockquote>Instead change the require of index.js in …/file.ts to a dynamic import() which is available in all CommonJS modules.</blockquote><p>This happens because you are trying to connect an ES module from the CommonJS module. Alas, in recent years, the number of such errors when updating packages and switching to new releases has been growing inexorably. Why do such errors occur? Let’s understand the nature of this phenomenon. To do this, you need to plunge into the history of JS.</p><h3>JS</h3><p>JS was born in 1995. Was written in a few months. It was initially loaded on the page via a tag script. All JS loaded on the page fell into the global scope. Because of this, when loading several external scripts, there were collisions, variable overloads, and other side effects. There were no imports and exports as such. Although keywords have been reserved in the language.</p><h3>Node.js and CommonJS</h3><p>By 2009, the technology had become so popular that there was a request why not use JS on the backend. Thus, Node.js appeared. Despite the similar syntax, it was arranged a little differently. Since, unlike the browser, run-time Node lacks html and the script tag, some other mechanism for loading JS was required to load the module. Such a mechanism was require. It’s worth noting that require is not a new language property. This mechanism was built into run-time.</p><p>Require.js parsed a module, built a syntax tree in it, found connections to other modules in it, chose the first one, connected it to the syntax tree, then parsed the next module. This process continued until the entire tree was built. Until this moment, the module was not considered loaded and it was not possible to analyze it. After parsing and building the tree, require exposed the module.exports object and provided an API to work with it, hiding the implementation details.</p><p>It is also worth noting that, unlike the browser, node.js could not search by URL, but searched by the path to the file, if the path was not provided, then the module was searched in the local node_module folder, if it was empty, then it was searched in the folder with the same name, but a level higher, and so on, until node.js reaches the root folder. This convention has been called CommonJS. To load external libraries, npm was created, which looked for a package in an external shared repository, or in those specified by the user, and pulled the necessary external libraries into the corresponding node_modules folders.</p><h3>Require.js and AMD, etc</h3><p>Some time after the release of Node.js and the appearance of the require mechanism, dudes from the world of browsers appreciated this approach and took steps to drag this approach to themselves. Porting 1in1 CommonJS to the browser is not possible. The first reason: in the work of JS within Node.js and the browser, is that the browser does not have access to the file system and operates on the url, while node.js operates on the file system. Reason two: run-time can add any variable to the script. Module.exports is not available in the browser. Therefore, in 2010, the RequireJS library was created. Which offered an adapted version of CommonJS for a browser called AMD. It is important to note that AMD and CommonJS did not change the language itself.</p><h3>ES6 and ESM:</h3><p>In 2015, ES6 appeared, which included many of the changes that the community asked for. Among other things, it described import and export. In 2017, these standards were implemented experimentally in version 8 of Node.js . They were ESM titles. The main difference between ESM and CommonJS was dynamic import, which allowed loading individual parts of the module as they were used. At the same time, the possibility of static import was preserved using the await import construct, which, in terms of implementation, resembles the implementation of require in CommonJS</p><h3>Chronology of events:</h3><p>Chronologically described events can be represented as follows.</p><p>1995:</p><ul><li><em>Browser (Script)</em></li></ul><p>2009:</p><ul><li>Node.js</li><li>CommonJS (Node.JS): Require | module.exports</li></ul><p>2010:</p><ul><li>AMD (Require.JS | Browser): Require | module.exports | Define</li></ul><p>2015:</p><ul><li>ES6 Standard: import | Export</li></ul><p>2017:</p><ul><li>ESM (Node.js 8.0.0 2017 Experimental | Node.js 12 2018): import | Export</li></ul><h3>Typescript</h3><p>Let’s add one more level of complexity. Historically, Typescript has been transpiled to CommonJS and has been a strict syntactic superset of JS. That is, it supported all the properties of the JS language, but at the same time added its own syntactic constructions to the language. It is important to note that natively browser and node do not support Typescript. To work with TS, you must first translate TS into a language understandable by node and browser. The second important point is that ESM support became available from April 8, 2022 for nodeResolution NodeNext.</p><p>Up to this point, TS could not:</p><ul><li>Create correct imports with extensions;</li><li>Parse exports and imports sections in package.json; Documentation link: <a href="https://www.typescriptlang.org/docs/handbook/esm-node.html#packagejson-exports-imports-and-self-referencing">https://www.typescriptlang.org/docs/handbook/esm-node.html#packagejson-exports-imports-and-self-referencing</a></li></ul><p>It is also important to note that jest still does not fully support ESM.</p><h3>Why the transition between CommonJS and ESM can be painful</h3><ol><li>It is very easy to import CommonJS packages into ESM. At the same time, reverse porting can be associated with difficulties.</li><li>Node.js was originally oriented towards CommonJS</li><li>There are a huge number of packages written in CommonJS. For example, all tooling (ts-node, typescript, etc), etc.</li><li>CommonJS made it possible to implement modularity in packages in different ways. As a consequence, different tools have different implementations of modularity.</li><li>CommonJS allowed external processes to modify node, for example, ts-node could modify node.js at runtime. Now this is impossible.</li><li>ESM imposes a certain kind of modularity, so a smooth transition is not possible.</li></ol><h3>Node.js recommendations for fresh packages:</h3><p>It is important to note that for new packages, ESM is the official standard format for packages to reuse code. Modules must be defined using import and export statements. Link to official documentation: <a href="https://nodejs.org/api/esm.html">https://nodejs.org/api/esm.html</a> Thus, we come to the conclusion that the transition is inevitable and unpleasant. What changes in Node.js make this transition more favorable at the moment? Experimental support for running ESM on the fly appeared in another 14 versions of Node.js, but only in version 18 did it become stable. In addition to this, the latest version of ts-node introduced on-the-fly transpilation.</p><h3>What have we done?</h3><p>Created a repository <a href="https://src.aligntech.com/projects/CCW/repos/ipl-shared/browse">https://src.aligntech.com/projects/CCW/repos/ipl-shared/browse</a> We created an ESM package in it on the latest JS stack: NodeJS 18. Checked launch. All OK. We tried to drag it into the infrastructure: Jest, Eslint. Added unit tests. Everything is working. We converted the package to Typescript and tried to run it through TS-node. Everything is working. There was a subtle point related to the fact that in the 18th version in the early time it is impossible to influence the work of its internal structure, so we had to write the following wrapper to run Typescript:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BP_naKhwe1H0dnBnDQqaMA.png" /></figure><p>But all this machinery worked only within the framework of a mono-repository. Why? Because for external packages, it was necessary to prepare a project either in the form of CommonJS or in the form of ESM.</p><p>Of course, we could solve all this through static imports const X = await import (‘Y’); but it’s not productive.</p><p>To do this, we set up the transpiler to generate esm and commonJS modules.</p><p>The following structure came out:</p><p>dist/</p><ul><li>esm/</li><li>commonJS/</li></ul><p>src/</p><ul><li>Package.json (where the module type was specified)</li></ul><p>In our situation, we can import a package from three points:</p><ol><li>from a neighbouring workspace via the ./src/index.ts folder</li></ol><p>2. when we import the transpiled code from another repository:</p><p>a. From ESM</p><p>b. From CommonJS</p><p>The scheme is working, but alas, it is not enough for full-fledged work. I will illustrate this thesis with an example:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4MsJDXO-pavWMqMM9pQ3UQ.png" /></figure><p>Of course, you can use createRequire and await import, but I would like to connect in the traditional way: Import { something } from “@aligntech-ipl/translations-core”; const { something } = require(“@aligntech-ipl/translations-core”);</p><h3>Why is it better not to do this?</h3><ul><li>Getting into the implementation details</li><li>Human factor as we add potential bugs in case the package structure changes</li></ul><h3>Conditional exports</h3><p>To simplify imports in code, we use two constructs in package.json:</p><ul><li>PublishConfig — this field contains an object with properties that will be replaced with properties with the same names after the package is built for publication</li></ul><p>It turns out like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/694/1*XwJP4iSxD7c0a1hirqLPdQ.png" /></figure><p>I’ll add some details to the screenshot:</p><ul><li>“.” is subpaths export. Thanks to him, we can implement different export conditions along the way</li><li>“types” — allow you to specify different types depending on the conditions</li><li>“default” is the generic entry point</li></ul><h3>What else did we have to do?</h3><ul><li>Even in the case of using conditional exports, we will have to change the js extension to mjs (cjs), either in ESM or in the CommonJS package.</li><li>By default js files will be interpreted by node.js depending on the value of the type field in package.json.</li><li>Additionally, we will need to replace the non-dual packages with their respective counterparts for the CommonJS and ESM versions of the package. For example, in ESM use lodash-es, and in commonJS lodash.</li></ul><h3>Summary:</h3><ul><li>We managed to set up ts-node launch</li><li>Transpilation on the fly turned out:</li><li>Acceleration of test launch compared to a similar launch by 6–10 times.</li><li>It turned out to set up dual-package (commonJS | esm)</li></ul><h3>Links:</h3><ul><li><a href="https://nodejs.org/docs/latest-v18.x/api/packages.html">https://nodejs.org/docs/latest-v18.x/api/packages.html</a></li><li><a href="https://medium.com/sungthecoder/javascript-module-module-loader-module-bundler-es6-module-confused-yet-6343510e7bde">https://medium.com/sungthecoder/javascript-module-module-loader-module-bundler-es6-module-confused-yet-6343510e7bde</a></li><li><a href="https://nodejs.org/api/packages.html#dual-commonjses-module-packages">https://nodejs.org/api/packages.html#dual-commonjses-module-packages</a></li><li><a href="https://nodejs.org/api/packages.html#conditional-exports">https://nodejs.org/api/packages.html#conditional-exports</a></li><li><a href="https://www.typescriptlang.org/docs/handbook/release-notes/typescript-4-7.html#packagejson-exports-imports-and-self-referencing">https://www.typescriptlang.org/docs/handbook/release-notes/typescript-4-7.html#packagejson-exports-imports-and-self-referencing</a></li><li><a href="https://2ality.com/2019/10/hybrid-npm-packages.html">https://2ality.com/2019/10/hybrid-npm-packages.html</a></li><li><a href="https://www.typescriptlang.org/docs/handbook/esm-node.html#packagejson-exports-imports-and-self-referencing">https://www.typescriptlang.org/docs/handbook/esm-node.html#packagejson-exports-imports-and-self-referencing</a></li></ul><h3>Level Up Coding</h3><p>Thanks for being a part of our community! Before you go:</p><ul><li>👏 Clap for the story and follow the author 👉</li><li>📰 View more content in the <a href="https://levelup.gitconnected.com/?utm_source=pub&amp;utm_medium=post">Level Up Coding publication</a></li><li>🔔 Follow us: <a href="https://twitter.com/gitconnected">Twitter</a> | <a href="https://www.linkedin.com/company/gitconnected">LinkedIn</a> | <a href="https://newsletter.levelup.dev/">Newsletter</a></li></ul><p>🚀👉 <a href="https://jobs.levelup.dev/talent/welcome?referral=true"><strong>Placing developers like you at top startups and tech companies</strong></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d6b62cfc851c" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/transpiling-typescript-into-double-packages-commonjs-esm-d6b62cfc851c">Transpiling Typescript into double packages (CommonJS + ESM)</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Development guidelines]]></title>
            <link>https://levelup.gitconnected.com/development-guidelines-d7381310ef06?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/d7381310ef06</guid>
            <category><![CDATA[development]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[hr]]></category>
            <category><![CDATA[talent-management]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Thu, 04 Nov 2021 04:17:41 GMT</pubDate>
            <atom:updated>2021-11-04T04:17:41.049Z</atom:updated>
            <content:encoded><![CDATA[<p>A quick introduction to Agile Mindset of Agile Developer</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MlDZ7EsZ_tOWjRwshVim7A.png" /></figure><h3>Introduction</h3><p>Onboarding new personel is a very crucial step. Imagine yourself as a newcomer. You walk into new territory, and everything is unknown to you. Yet, you are chosen for the task and capable of doing your work. But what amount of time and you will waste your inner resources if somebody won’t guide you into your new line of work? What costs you and the company you are working for will be a total lack of information on how you should do your job.<br>Every company is different, and every culture has its differences. And to describe an example of how our company could solve things, I want to share with you the following guidelines.<br>Our company vision and development guides from our department form the following PRINCIPLES, STANDARDS, and METHODOLOGIES. We use them in our work daily.</p><h3>Principles</h3><p>We base our work on the following principles of Agile Development:</p><h4>SOLID concepts</h4><ol><li>The <a href="https://en.wikipedia.org/wiki/Single-responsibility_principle">Single-responsibility principle</a>: “There should never be more than one reason for a <a href="https://en.wikipedia.org/wiki/Class_(computer_programming)">class</a> to change.” In other words, every class should have only one responsibility.</li><li>The <a href="https://en.wikipedia.org/wiki/Open%E2%80%93closed_principle">Open–closed principle</a>: “Software entities … should be open for extension, but closed for modification.”</li><li>The <a href="https://en.wikipedia.org/wiki/Liskov_substitution_principle">Liskov substitution principle</a>: “Functions that use pointers or references to base classes must be able to use objects of derived classes without knowing it.” See also <a href="https://en.wikipedia.org/wiki/Design_by_contract">design by contract</a>.</li><li>The <a href="https://en.wikipedia.org/wiki/Interface_segregation_principle">Interface segregation principle</a>: “Many client-specific interfaces are better than one general-purpose interface.”</li><li>The <a href="https://en.wikipedia.org/wiki/Dependency_inversion_principle">Dependency inversion principle</a>: “Depend upon abstractions, [not] concretions.”</li></ol><h4>“Don’t repeat yourself”</h4><p>(DRY, or sometimes “do not repeat yourself”) is a <a href="https://en.wikipedia.org/wiki/Principle#Principle_as_axiom_or_logical_fundament">principle</a> of <a href="https://en.wikipedia.org/wiki/Software_development_process">software development</a> aimed at reducing repetition of software patterns, replacing it with abstractions or using <a href="https://en.wikipedia.org/wiki/Data_normalization">data normalization</a> to avoid redundancy.</p><h4>Code reuse</h4><p>Code reuse, also called software reuse, uses existing <a href="https://en.wikipedia.org/wiki/Software">software</a>, or software knowledge, to build new software.</p><p>We are trying to use tools and libraries like “lodash” to speed up development and avoid re-inventing the wheel.</p><h4>Package principles</h4><p>Organize classes in more extensive systems to make them more organized and manageable.</p><p>In our department, we are highly encouraged to use great Yarn 2.0 features like “workspaces.” Our department is responsible for more than six applications, and we split them between 3 repositories. They use the same API and work with the same date. To reduce code duplication, we use the internal package system and Yarn 2.0 ’s workplaces. This way provides us with some middle ground before the private publication of packages.</p><h4>“You aren’t gonna need it” (YAGNI)</h4><p>is a principle of <a href="https://en.wikipedia.org/wiki/Extreme_programming">extreme programming</a> (XP) that states a <a href="https://en.wikipedia.org/wiki/Programmer">programmer</a> should not add functionality until deemed necessary.</p><p>In our practice, we are trying not to force things. For example, on the first implementation, we might not know how customers will use this functionality. So we are trying to implement the most common and obvious ways possible. This approach allows us to show implementation ASAP (As Soon As Possible) and confirm our offering matches needs.</p><h3>Standards</h3><p>In our work, we try our best to follow industry standards:</p><ul><li>W3C;</li><li>ES6;</li><li>Typescript guidelines;</li><li>WebGL.</li></ul><p>To lower code review labor and provide some automation, we are using the following instruments of static code analysis:</p><h4>Linting</h4><p>Our team stores ESlint and Typechecking rules config as a package, which all our teams are expected to use. We also allow the possibility to tune local settings for each group.</p><h4>Sonar rules</h4><p>We use this tool as an instrument for the visualisation and storage of static code analysis. For our project, we use default Quality Gates. It is based on the “Fix the water leak concept.” More info here: <a href="https://docs.sonarqube.org/7.4/user-guide/fixing-the-water-leak/">https://docs.sonarqube.org/7.4/user-guide/fixing-the-water-leak/</a></p><h3>Methodology</h3><p>Methodology is a blueprint of the process of development. It’s necessary step which leads to</p><h4>Prototyping UI (User Interface)</h4><p>We are highly encouraging our developers to prototype UI tasks before starting implementation. This easy step allows us to discuss future implementation with stakeholders and validate the vision. If possible, we are highly encouraged to delegate this task to our developers or implement UI together with them.</p><h4>Mocking backend data.</h4><p>To prevent delays, we ask our developers to negotiate the implementation of backend requests with the backend developer. It allows us to form a blueprint, which will serve as a contract for future data exchange. We highly recommend adding these requests to a mocked server. In our case, we use Service Worker, which captures network requests and replies with mocked data.</p><h4>Writing tests</h4><p>I highly recommend that you provide yourself with tests before starting implementation. It will be easier to control the quality and manage the completeness of the code. We can divide Our unit tests could into four groups:</p><ol><li>UI. Primarily we test that the interface handles data sets well. For convenience, we use snapshot testing. Then we add tests that fire required events to trigger necessary parts of code.</li><li>Tools. Then we test helper functions. We feed them various data and check that they handle data correctly.</li><li>Backend. We use nock. It allows us to update requests and mock requests to the backend easily.</li><li>Store. In our case, our application uses Redux. In this case, we must check thunks actions. It is easier to provide default state for part of Store, trigger required actions and compare actual movements with the standard set of steps.</li></ol><h4>Development</h4><p>Don’t rush things — split tasks into smaller chunks. Implement a happy path first. Use stubs. Don’t try to implement everything from one go. Save often!</p><h4>Code review</h4><p>When you accept:</p><ol><li>Be grateful and polite. Remember that another person helps you.</li><li>Try to understand the other person’s view even if you think that he could be wrong.</li><li>If feedback is optional or it would take more time to implement that current temporary solution. It is a beautiful tradition to create a tech debt task, which the developer will implement later.</li><li>If something is unclear, don’t throw debates. Make a call and share the screen. It’s easier to understand when you are both looking at the same thing.</li></ol><p>When you review someone:</p><ul><li>Try to be direct, confident, analytical, and friendly.</li><li>Try to be specific and provide small code examples of your idea. We are not playing riddles or trivia.</li><li>Do not forget to praise for intention. We are on the same boat, and we are pushing our company towards the grand goal!</li><li>Suppose you want to judge another person’s work. Please note that people are susceptible to critics. So please be very cautious while selecting three options:</li></ul><ol><li>Block merge (Highly conflicting option) .It’s when for some reason, it’s not great to merge PR. Please provide some comments on why you are doing this.</li><li>Task (Moderate conflicting option). Suppose this is an issue that the developer should fix before merging. Tasks are greater than comments: they have links to code versions, and developers could check their PR progress.</li><li>Comment — If you have an idea and you think that it is optional. It is excellent to place a comment for a colleague.</li></ol><h4>Validation</h4><p>It is a process when we try to determine that our functionality works as expected before handling it to the SQA engineer. It’s a very crucial step — it raises trust in your work, and you can ruin it easily by not checking some visible parts in your work. This process should be performed in two places:</p><ol><li>Locally. We could check all implemented steps on local env. It is easier to debug, but data could be outdated.</li><li>On Environment. We deploy our functionality via Bamboo on our testing ENVs. Then we check all requirements described by our analytic.</li></ol><h3>Conclusion</h3><ol><li>People from outside are not familiar with your procedures, so it’s vital to provide some information on how you expect them to do their job.</li><li>The guideline is also a vital step for your company’s transformation because it’s easier to transform a newcomer and show their behavior as an example of how your colleagues should do things.</li><li>Guidelines save a lot of time, nerves, and effort because they show a valid path.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d7381310ef06" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/development-guidelines-d7381310ef06">Development guidelines</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to manage technical changes]]></title>
            <link>https://levelup.gitconnected.com/how-to-manage-technical-changes-fd74c3f647d6?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/fd74c3f647d6</guid>
            <category><![CDATA[typescript]]></category>
            <category><![CDATA[development]]></category>
            <category><![CDATA[react]]></category>
            <category><![CDATA[typescript-with-react]]></category>
            <category><![CDATA[wysiwyg]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Tue, 02 Nov 2021 02:36:50 GMT</pubDate>
            <atom:updated>2021-11-02T12:46:29.013Z</atom:updated>
            <content:encoded><![CDATA[<p>Based on Rich Text Editor selection</p><h4>Description of the business task:</h4><p>To support one of our business processes, we selected a rich text editor. Which should fulfill basic requirements:</p><ul><li>Line formatting (headers, paragraphs)</li><li>Symbol formatting (italic, bold, underscore, colorizing)</li><li>Saving state on the backend</li></ul><p>And two specific requirements:</p><ul><li>It must have end-to-end numbering, and each item must have its checkbox;</li><li>The editor should indicate the progress of completion in the toolbar of the editor.</li></ul><p>Our stack for frontend is:</p><ul><li>React</li><li>Redux</li><li>Typescript</li></ul><p>The first version of this editor was based on Draft.JS. Our former developer (Vladimir Ivanov) implemented it more than a year ago. And it seemed like a great choice:</p><ul><li>Facebook supported it</li><li>It worked great with our stack: React and Typescript</li><li>It had all the formatting options we need</li></ul><p>And it did an excellent job for a year.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/736/1*kUGW3JmZagMeiWF94Y5UhA.png" /><figcaption>Draft.JS implementation</figcaption></figure><h4>Until we found out that it had three flaws:</h4><ul><li>Facebook developers didn’t update it for one year;</li><li>Its supporters don’t allow incoming PRs;</li><li>It had a minor issue with the incorrect positioning of the anchor when the user clicked on the rightmost part behind the end of the text. It placed anchor at the start of the line. Not at the end.</li></ul><p>It was crucial for our stakeholders, so I was assigned to fix this issue. After research, I found two solutions:</p><ul><li>Check that user clicked behind the end of the line. Then force anchor at the end of the line. It worked great, but only if the text block wasn’t multiline. If it was multiline, then it placed anchor at the end of the last line;</li><li>The other solution was to calculate the cursor’s position, calculate the height of the text block, calculate the number of rows, calculate the same row, and then calculate its width. Then check if the user placed the click before or after the end of the line. It worked pretty well, but with a noticeable delay.</li></ul><p>Both decisions seemed like a crutch.</p><h4>Alternatives</h4><p>We decided to look for a replacement. As you know, there are a ton of Rich Text Editors. I will name only the most popular ones:</p><ul><li>Quill</li><li>Slate</li><li>TinyMCE</li><li>Draft.js</li></ul><p>So I decided to investigate the matter seriously. First, I looked through various reviews and articles.</p><p>After this, I decided to check <a href="https://www.npmtrends.com/quill-vs-slate-vs-tinymce-vs-draft-js">npm trends</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9As3DwM6MpSgiUQGM3MWlw.png" /><figcaption>Comparison of npm trends</figcaption></figure><p>And it seemed to me that my findings pushed me to choose an editor that I had already thrown away.</p><p>After long hours of research, I decided to prototype two versions of editors: Quill and Slate. I decided to drop TinyMCE and Draft.JS because some of our team members had expertise with them. I implemented a proof of concepts, but after some thought, I decided that it wasn’t enough to make my mind.</p><h4>Decisions through democracy and teamwork</h4><p>My doubts were based on the fact that I don’t work solo. I work with the professional team, which will work with this code too. It will fix bugs, maintain them and add new features. So I can’t make this decision alone.</p><p>I decided to do an experimental thing. I gathered a list of things left to do, created JIRA tasks, and asked my coworkers to take equal parts of this work among themselves. So they will gain expertise with each editor, and then I will ask them to vote for editors.</p><ol><li>How to find out about the expertise of the product from the participants? I conducted a survey, and the score dropped to know/don’t know.</li><li>In this situation, we limited ourselves to solving one question: “What are we replacing?”</li><li>The distribution of tasks was democratic. Team members took on the jobs that were convenient for them. The only limitation was that everyone had to work with an unknown product/approach/library.</li><li>The decomposition work requires deep immersion in the project.</li></ol><h4>Voting</h4><p>After one week of work, I created a Poll. In it, I asked to select any number of editors. We were choosing between Quill, Slate, Draft.JS, and something else.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6y0XinaqffmxvGeI6Kr5Mw.png" /></figure><h4>Subjective assessment of this approach:</h4><p><em>Pros:</em></p><ol><li>This approach generates minor disagreement in the future, as it implements the principle of collective responsibility;</li><li>This path leads to increased expertise in the team and reduces the bus factor.</li><li>Discussions are most fruitful as all participants have a similar level of engagement.</li></ol><p><em>Cons:</em></p><ol><li>This path is the most expensive in terms of experimenting;</li><li>It is poorly compatible with situations where an urgent decision needs to be made.</li></ol><p><em>Summary:</em></p><ol><li>In my opinion, this approach is beneficial in situations where further productivity and agreement are of more excellent value than the cost of experimenting — for example, choosing a framework or selecting a storage system for a project.</li><li>This approach is based on finding a mutually beneficial position. It’s called a Win-Win scenario. You can read more <a href="https://www.mindtools.com/CommSkll/NegotiationSkills.htm#:~:text=A%20win%2Dwin%20negotiation%20is,then%20that&#39;s%20a%20win%2Dwin!">here</a></li></ol><h4>Feedback from the team</h4><p>Here are their thoughts:</p><blockquote>The following were some pros/cons of voting for Slate:</blockquote><blockquote>Plus:</blockquote><blockquote>- react out-of-the-box — hooks, components, easy/familiar to work with</blockquote><blockquote>- programmatically flexible, customizable</blockquote><blockquote>- documentation — good, easy to follow, good examples</blockquote><blockquote>Minus</blockquote><blockquote>- still beta</blockquote><blockquote>Quill — at first glance, it seemed that this was a very convenient library; when I opened the documentation, everything seemed logical and understandable. However, the first problem I encountered was the weak connection with the reactant; for example, when customizing the toolbar, the Developer cannot pass the react component and needs to work directly with the DOM tree. Also, although the documentation is quite clearly described, many different options are not marked in it, and after reading it, you will not be able to understand how to solve a specific problem that is quite slightly beyond the default.<br>You have to look for issues in the git and look at quill implementations in other projects to understand how to implement this correctly. For example, when creating a progress bar, the Developer spent much effort finding a data format supported by the quill. It turned out that you can write many things differently from the documentation; it is not reflected in it.</blockquote><blockquote>Slate — it is immediately apparent compared to quill that it is tailored for React, the documentation covers more different use cases, and slate also has a live community in slack where you can turn for advice. However, the code in slate itself looks more complicated. Take the standard progress bar, the code they already offer — you can write this yourself for a very long time; incomplete type coverage is a little misleading, and sometimes ts tells you that there are no properties, although they are. In general, since slate works very harmoniously with react, a large number of working examples, and the live community, slate bypassed quill for convenience</blockquote><blockquote>Draft.js is a reasonably popular editor. However, it is not very flexible for custom formatting. But it suits pretty well if there is a need to use it as a regular text editor. It has a complex data structure for saving data from the editor. Separate formatting intervals (bold text, for example) are stored in numbers (which character to apply from and the length of the interval). Badly updates formatting if you replace text via JS functions, i.e., automatically.</blockquote><blockquote>Quill. In my opinion, the most straightforward editor to write code. Not very flexible, but very convenient as an editor without custom formatting. For me, the inconvenience is that there is no separate component for the formatting buttons (text color, text format, and so on). But on the other hand, it speeds up the writing of code. Quill has a convenient form for saving data. Each fragment has its format, and there is no transfer between the fragments, which text he gave to him; he will display this. The hyphenation must be written with special characters.</blockquote><blockquote>Slate. You have to write much code in a well-customizable editor, even if you need a short version with minimum functions. The data storage object is very similar to Quill, and I find it quite convenient. A Slate framework is an excellent option since we control all editor functions and can create our own.</blockquote><h4>Implementation</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/742/1*s8uL6boq9idJAkDy-dS8FA.png" /><figcaption>Final Slate implementation</figcaption></figure><h4>Conclusion:</h4><ol><li>Don’t base your decision <em>only</em> on numbers.</li><li>If I had chosen the usual way (through trends and articles), I would select an editor that would not suit me.</li><li>If I had stopped and didn’t share my choice with the team, I would have chosen Quill, an editor that was easier to set up but ultimately disliked due to problems with customization.</li><li>In my opinion, the manager should not force his decision. The team should select working tools from possible variants.</li><li>This decision will come easy, won’t cause any protest, and will enlarge the leader’s image.</li><li>Be friendly and kind to your team members. A good relationship is a win-win long-term investment.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fd74c3f647d6" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/how-to-manage-technical-changes-fd74c3f647d6">How to manage technical changes</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Organization of the development department from scratch]]></title>
            <link>https://maxbogus.medium.com/organization-of-the-development-department-from-scratch-eba0aa5ef5c5?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/eba0aa5ef5c5</guid>
            <category><![CDATA[development]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[teamwork]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[management]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Wed, 20 Nov 2019 18:23:05 GMT</pubDate>
            <atom:updated>2019-11-20T18:23:05.135Z</atom:updated>
            <content:encoded><![CDATA[<h4>I tried to make a squeeze out of my observations and get conclusions from them. I want to warn you: the topic is extensive, so I’ll tell you the essential things. If you have different questions, doubts, and turbulence; as a result, that’s good. Ask me them by mail or through social networks.</h4><h4>I base my post on more than ten years of experience in organizing various types of activities and groups:</h4><ul><li>As a foreman of the brigade of installers, I organized networking activities;</li><li>As a teacher, I held the events of students to achieve the goal of pursuing the course;</li><li>As a customer, I organized the work of contractors;</li><li>As the head of the testing department, I held a job to achieve the required level of quality of projects in the development.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XkwemZ4l_Fwp1T50JzhbaA.jpeg" /><figcaption>Flying Cloud — Some ships of the clipper ship era, their builders, owners, and captains; (1913)</figcaption></figure><h3>Define the concepts</h3><p>The organization is a process, an activity to create and improve the interrelations between parts and elements to bring order to operations and improve their effectiveness.</p><p>No matter what activity — the principles of organization are the same. Yes, each group has its characteristics (specificity of events, culture, ways of implementation), which the head takes into account. Still, the essence of the organization as an activity does not change.</p><h3>Why is it so difficult to start from scratch?</h3><p>On an established project, much has already been determined. Previous managers already established processes and hired people. We are already in the process and take into account the circumstances.</p><p>At the start, you start from scratch. Eyes run-up, you have to do so much, and it’s unclear how to start, where to put priorities, what tasks to choose. At this stage, you are already waiting for results from the program, but nothing is ready yet.</p><h3>The first relief:</h3><p>In fact, with all the problems of growth and development, it’s easier to grow. Since energy is spent less than when stabilizing and maintaining a balance. But there is a high share of uncertainty and risks.</p><h3>Where do we start?</h3><p>First, we need to understand where to develop. You need a clear understanding of the goal to know more precisely where you will move and how you will realize that you are going in the right direction? The purpose defines the criteria for achieving.</p><h3>What’s next?</h3><p>To determine the goal, talk with your manager and the heads of departments with whom you will work. Find out what they expect from you. Have your manager expect these expectations.</p><p>For navigation in space, it is crucial to understand where you are right now. Ask the following questions to yourself:</p><ul><li>What means do you have?</li><li>What do you have?</li><li>What is the budget?</li><li>What means of production?</li><li>How many people are there?</li><li>What time frame?</li></ul><h3>How to move towards the goal?</h3><p>The primary way to move to the goal is to control tasks. As soon as you begin to understand the path, you can estimate that with the current means for you is realizable and what is not.</p><p>Remember the project triangle? Money, time, demands, and in the middle quality — it’s the central control point of a project. You can ask for more resources, reduce the variety of functions, throw out jobs, ask for more time.</p><p>Restrictions are an essential indicator. They allow you to focus on reality and execute a project. They give definiteness, they gather people into a pile and will enable them to do something. Without these frames, you will be like an amorphous amoeba.</p><p>What is the power of limitations? They give control points. Your task as a leader is to feel these limitations. To look at them in terms of their importance and understanding of what you can do with them.</p><h3>You discussed the tasks, discussed the stages</h3><p>Act iteratively. For project management, many use the Agile methodology. Why did they take root so? They make it possible to bring together the leadership and subordinates, make the processes more transparent, and allow receiving feedback.</p><p>Why is it important? You create a zone of trust, but do not lose control of the project. You allow people to choose things (solutions), but at the same time lower them tasks from above and watch out for the result.</p><p>The choice of methodology does not affect management. There are basic principles that exist in any production. It does not matter the industry, standards, people, and stuff. These nuances, of course, change the appearance and make adjustments to the processes, but people are still people.</p><h3>When is the formation of the department?</h3><p>After you implemented the formation of the department occurs as projects. Do not immediately describe in detail every millimeter of the path. So you will not do anything. Excessive planning drowns projects. Ideally, if the process is fixed and detailed, it is necessary if necessary. As you enter new people, you:</p><ul><li>fixing arrangements;</li><li>fix and approve plans;</li><li>Establish job duties and regulations.</li></ul><p>It is worthwhile to think about the formation of the department as the portfolio of projects grows. While you only have one project, you only need to pull it. You are still too weak to formalize processes and recruit people.</p><p>On the other hand, planning, regulations, and job descriptions create a framework and allow for the recording of agreements. They make it possible to reduce the zone of confusion in the minds of people.</p><h3>Matrix vs. Hierarchy vs. Agile</h3><p>The organization depends on the context, the customer, the specifics of the product, but it does not change much. In our case (the Gett company) on the face is a combination. When there is a matrix in the hierarchical structure in which there is an agile of a team. Whose members are subordinate to the development manager who submits to the director.</p><p>It’s like burning a fire — balance is essential: too much wood — and you can not light a fire, it’s not enough — the fire will quickly die out and give a little heat.</p><p>I proceed from the idea that it is better to lack and shortage, as this gives mobility and clarity. On the other hand, this approach deprives you of reserves that can be useful on a rainy day.</p><h3>Stages of department growth: Employee.</h3><p>While you are alone, maybe you do not need more documentation and everything else. You would have the project done. Ask people only in case of flooding with tasks. The lack of functions is worse than their excess. When there are not enough tasks, productivity is lost, people begin to engage in nonsense. They believe that you paid them for doing nothing, and when work comes, it is difficult for them to return to their previous channels.</p><p>When to recruit people? My recipe is this: If you work more than 2–3 months — you need an assistant. When you work for 1–1.5 weeks, people should not be taken.</p><h3>Stages of department growth: Team.</h3><p>The team begins to form after 3–4 people. Then the leader is revealed. But remember that the more people, the more difficult it is to negotiate. What two people agreed for 1 hour, 3–4 people will agree 1–2 days. Contact is already lost, conflicts, and misunderstandings arise. The more people, the less you will program, and the more you will communicate.</p><p>When you have more than six people under you, you can forget about programming in general. Up to 6 people, you most likely will not require testers, designers, analysts. All this you will manage to do by the team.</p><p>At the stage of 6–7 people, there is a need to formalize the processes, clarify the agreements, introduce regulations. Why is this necessary? Very simple. To protect the vocal cords. When you start telling the same thing from time to time, and people will ask the same questions — it means that it’s time for regulations, and you need to start recording processes. Do not be afraid to make a mistake. You will be told immediately about the poorly developed methods.</p><h3>Stages of department growth: Department.</h3><p>The department is a structure in which there are more than 6–7 people. There may already be several teams in it. As you grow, there is a need to distribute work among people. Perhaps you want to make dedicated specializations.</p><p>These specializations will be needed when you have too many profile tasks, and you will need to allocate the resource to reduce the non-core load from the team: drawing the design, interface creeping, working out requirements, working with feedback from users.</p><p>You will need helpers in the leadership. Remember that effectively. The leader manages 5–7 subordinates. If more people, he begins to lose the focus of attention.</p><h3>The life cycle of the team</h3><p>We can distinguish five stages of the life cycle of the department:</p><ul><li>Origin. Only the idea arose, and the first person appeared.</li><li>Growth. The team is growing. The main thing here is not to stop the process. But do not give the team too much growth.</li><li>Stability. The team is stabilizing. Be careful. You can lose the tempo. The group periodically needs to throw in new Challenges.</li><li>Stagnation. Search for what you drag down. Change. Solve problems.</li><li>Death. The command is bent. What to do? Update the management style. Change the team or leave.</li></ul><h3>Management Position</h3><p>You are a leader, but neither a friend, nor a brother, nor an enemy. You are a person who is higher, has resources and power.</p><p>Your position is to create conditions for the group to be able to move unhindered to the goal. All you have to do is identify problems, assess risks, make decisions, and facilitate them in a group.</p><h3>Problems:</h3><p>In my opinion, in the process of organizing the department, the issues are as follows:</p><h3>Loss of control</h3><p>Loss of control is a normal situation. You can not control everyone. And as power grows, control is lost. It is crucial to organize a structure that will stabilize itself. With your participation, of course. The more you control the structure, the weaker it becomes. Remember about 5–7 focuses of attention? That’s the level and will be the performance of your organization if you keep everything under control.</p><h3>Informal leaders and opposition</h3><p>…This is normal. Each team has its chat room, where the dice are washed (first of all, to you). There is always someone dissatisfied with you. There are still people who do not like something. If they start spoiling everything, then get rid of them by any legal and available means. With haters, the question is subtle, if it does not destroy you, and you do not lose credibility — leave. You don’t need to show cruelty and tyranny. It is crucial to be able to listen and hear them. When they are not pressed and listened to, they lose influence over the situation. And they can give valuable ideas for development that will allow you to achieve great success.</p><h3>Riots</h3><p>When your activity does not bring satisfaction, and the team is already ready or going against you.</p><p>The very trap is the loss of the head in the process of work. And illusions are the worst thing. When you are wrong, and there is a crisis, you face a harsh reality, and, in fact, reality itself gives you feedback and limits you in choosing. Worse, when everything is going fine, or you think everything is going fine.</p><p>Because in this situation, you do not control anything. It is essential at the same time not to forget about the central question: Am I not doing shit?</p><h3>Assholes and psychopaths</h3><p>These are people from whom you need to get rid of immediately. Usually, they smile at you, and people behind you set people against you and abuse others.</p><p>Who are they?</p><p>It is necessary to distinguish simply people with a bad mood and habits from pathologically ill.</p><ul><li>Assholes:</li><li>Maybe a person has a bad mood</li><li>People with pathologies:</li><li>Psychopaths — a pathological disruption of the brain</li><li>Sociopaths — antisocial worldview and behavior</li></ul><h3>What prevention can be:</h3><ul><li>It is necessary to find out the situation, analyze the connections between people, the mood in the team, the influence of the person on decisions and motivation</li><li>Track dismissals and loss of productivity</li><li>Carry out face-to-face bets, analyze.</li><li>Fix arrangements</li><li>Regularly monitor performance</li><li>Reduce emotional communication</li><li>Speak with the surrounding decisions</li><li>Do not be an asshole</li></ul><h3>Conclusions</h3><ul><li>Winners are not judged. Do not be afraid to make mistakes. If you succeed, management will forgive for your mistakes.</li><li>Work with management — these are those who give you resources, forgive your mistakes, and deliver new projects.</li><li>Do not strive for empty power. Empty power is when you have opportunities, but there is no purpose. Such power is easy to lose.</li><li>Trust, but in case of problems, fix the arrangements. Write meeting minutes.</li><li>The most terrible thing is the courts and labor inspections. Do not enter into a deal with a conscience — do not wash.</li><li>The organizational form of the department will cease to exist after your departure — this is the norm. Please do not make a monument out of him.</li><li>Work should not be the only joy in your life.</li></ul><h3>Literature</h3><p>I want to advise you some research, that would develop the topic of the post:</p><ul><li><a href="https://www.amazon.com/Cognition-Bradford-Books-Edwin-Hutchins/dp/0262082314/">Cognition in the wild</a> — a book of the cognitive psychologist about how to organize the administration of a military ship (crew management, navigation, team psychology). It’s no secret that software development is very similar to construction and shipping. Therefore, this book will help you understand the concepts.</li><li><a href="https://www.amazon.com/Being-Geek-Software-Developers-Handbook/dp/0596155409">Being geek</a> is the book of the Netscape development manager, which passed from engineer to manager and described this experience. The book tells different stages of growth (interview, communication with colleagues, transfer from engineer to management).</li><li><a href="https://www.amazon.com/Managing-Power-Politics-Influence-Organizations/dp/0875844405/ref=sr_1_1?s=books&amp;ie=UTF8&amp;qid=1508711806&amp;sr=1-1&amp;keywords=Jeffrey+Pfeffer%2C+Power%2C+influence+and+politics+in+organizations">Jeffrey Pfeffer, Power, influence, and politics in organizations</a> is a very cool book about what management, politics, authority, and power are. About how they are received, how they are lost, and what to do about it.</li><li><a href="https://www.amazon.com/Ethnogenesis-Biosphere-Leo-Gumil%C3%ABv/dp/5010020106/ref=sr_1_1?s=books&amp;ie=UTF8&amp;qid=1508711919&amp;sr=1-1&amp;keywords=lev+gumilev">Ethnogenesis and the Biosphere, Lev Gumilev</a> — a book by a famous researcher, is a story about how nations were born. Who are passionate individuals, how the group develops, what stages of life it has, and connection inside group?</li><li><a href="https://www.amazon.com/Snakes-Suits-When-Psychopaths-Work/dp/0061147893">Snakes in Suites</a> — a book about psychopaths and sociopaths in the work environment. About why they are many in modern companies and backgrounds. About what they are dangerous, how to identify them, screen out, and what to do if you work with one of them.</li><li><a href="https://www.amazon.com/Asshole-Rule-Civilized-Workplace-Surviving/dp/0446526568/ref=pd_bbs_sr_1/002-1591139-6524826?ie=UTF8&amp;s=books&amp;qid=1173730032&amp;sr=1-1&amp;tag=jezebelamzn-20&amp;ascsubtag=9c7c8a917810821e128b248ce57d263740ed0f27&amp;rawdata=%5Br%7Chttps%3A%2F%2Fwww.google.ru%2F%5Bks%7Cgoogle%5Bt%7Clink%5Bp%7C1671412342%5Ba%7C0446526568%5Bau%7C5820775451778960292%5Bb%7Cjezebel">The No Asshole Rule: Building a Civilized Workplace and Surviving One That Isn’t</a> — a book about how to build relationships, about what hairy assholes and unpleasant communication. What to do about it?</li></ul><h3>Based on my speech</h3><p>Here is <a href="https://www.slideshare.net/maxbogus/organization-of-development-from-scratch-what-to-start-with">link</a> to slides.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eba0aa5ef5c5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Sleeping tips, or how to get up early?]]></title>
            <link>https://maxbogus.medium.com/sleeping-tips-or-how-to-get-up-early-6a816c56cf09?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/6a816c56cf09</guid>
            <category><![CDATA[life]]></category>
            <category><![CDATA[health]]></category>
            <category><![CDATA[sleep]]></category>
            <category><![CDATA[life-hacking]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Wed, 20 Nov 2019 15:17:05 GMT</pubDate>
            <atom:updated>2019-11-21T10:20:49.680Z</atom:updated>
            <content:encoded><![CDATA[<h3>How to get better sleep and wake up early?</h3><h4>A casual conversation with a colleague gave me an idea to write a compilation of my observations and knowledge that once helped me at the time to move to a stable sleep.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vuLaBaMJsKmuWc44QiudHA.jpeg" /><figcaption>from <a href="https://www.flickr.com/photos/s0ulsurfing/2759585402">https://www.flickr.com/photos/s0ulsurfing/2759585402</a></figcaption></figure><h4>Why is healthy sleep so important?</h4><ul><li>During sound sleep in the night-time period (from 23:00 to 4:00), melatonin is produced in the human body. Its absence can lead to type 2 diabetes.</li><li>During sleep, the brain undergoes the neural restructuring process, leading to the assimilation of knowledge and processing of emotional situations.</li><li>During the 8-hour sleep, a healthy person burns about 1500 calories. Lack of sleep causes storing unburnt calories and weight gain.</li><li>A person who has a lack of sleep is experiencing a shortage of energy, which is compensated by increased appetite, also leading to fattening.</li><li>Abnormal sleep can cause bipolar disorder.</li></ul><p>It is worth remembering that sleeping needs are changing over the years. There’s been a lot of research and debates on the subject. I adhere to the idea that it is better to sleep for 8 hours. However, I came across a study in which researchers examined the need for sleep at different ages. Here are their results:</p><ul><li>20–30 years: 7.23 hours.</li><li>40–55 years: 6.83 hours.</li><li>66–83 years: 6.51 hours.</li></ul><p>After I clarified the importance of healthy sleep, it is time now to talk about how to get enough sleep and come to work on time. It is no secret that to get used to that and to have a good sleep — you need to go to bed early. So we need to calculate time to find out when to go to sleep and when to wake up so that it is comfortable for us to get ready to go. To do this, you need to calculate:</p><ul><li>Time spent on the road. For example, for 1 hour.</li><li>Packing time (shower, breakfast, water, gymnastics). For example, 30 minutes.</li><li>Sleep time. For example, 8 hours.</li></ul><p>It turns out that to come work at 9:00, you need to go at 23:30.</p><p>When a person is tired and had a lack of sleep, he or she can go to sleep smoothly. But what if you do not want to sleep. How to help yourself? Firstly, try to secure comfortable sleeping conditions:</p><ul><li>Fresh linen;</li><li>Access of the fresh air;</li><li>The average level of humidity in the room;</li><li>Try to drink at night a glass of clean water, as dry mouth and thirst interfere with sleep. Water must be plain (tea and coffee are not suitable).</li></ul><h4>To make it easier to go to sleep, you need to consider several factors:</h4><ul><li>Watching something bright and emotional before going to bed is not recommended.</li><li>If you are overwhelmed by thoughts and plans, it is better to record and leave for the next day.</li><li>Refrain from smoking for 1–2 hours before bedtime.</li><li>It is better to give up alcohol a few hours before bedtime. On the one hand, it will be easier to fall asleep, but on the other, you won’t reach the deep sleep phases, and therefore, won’t get enough sleep.</li><li>Be careful with tea and coffee before going to bed. Tea has up to 6–8 hours of stimulation level. Coffee, if it is a good one — 30 minutes of stimulation, then 2 hours of inhibition. In both cases, there is dehydration, which may hinder sleep.</li><li>Stop eating 3–4 hours before bedtime. Since food causes heaviness and gives a certain amount of stamina, it may be difficult to sleep.</li><li>It is perfect if you are physically tired. It is recommended to make 10,000 steps a day.</li></ul><p>If you experience any problems, you should analyze the issues with sleep. Is there anything that can help you? Today there are a lot of applications and devices for sleep logging. They help to gather data and understand why you have a lack of sleep. Information you should collect:</p><ul><li>Level of CO2 / O — the less oxygen and more carbon dioxide are in the room, the worse your rest will be.</li><li>Noise level — noises attract our attention and take us to a light sleep phase.</li><li>Level of illuminance — bright light activates the nervous system and takes people to a light sleep phase.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6a816c56cf09" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Benefits of transferring TypeScript types to DefinitelyTyped]]></title>
            <link>https://levelup.gitconnected.com/benefits-of-transferring-types-to-definitelytyped-20b45141554?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/20b45141554</guid>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[npm]]></category>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[typescript]]></category>
            <category><![CDATA[definitelytyped]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Tue, 19 Nov 2019 13:43:26 GMT</pubDate>
            <atom:updated>2019-11-20T14:25:22.736Z</atom:updated>
            <content:encoded><![CDATA[<p>We had 32 custom typings for JavaScript libraries in our repository. At first, it started as a small pull request to a Webpack plugin package. After realizing that this process could save us time in the future, I proposed we move typings from our repository and use them from NPM.</p><h4>The planned process was the following:</h4><ol><li>Fork the DefinitelyTyped repo and create a branch for every package.</li><li>Copy our definition and check that everything works as it should.</li><li>Push and create a pull request.</li></ol><p>At first, it seemed that we could do this in one week.</p><h4>But I faced several difficulties:</h4><ol><li>Not every definition in our repo complied with DefinitelyTyped guidelines. So I had to adapt our definition to the guidelines first. Sometimes I had to rework the whole definition completely. But in the end, I got 32 fully described definitions.</li><li>We didn’t have tests for our definitions. It was a strict requirement for every pull request created into the repository. So I had to create tests for them. At first, I started creating test definitions without copying the use cases to tests or applying use cases to the existing definition. This prolonged transfer process: sometimes, my definition didn’t match its use case. After this, I started to copy use cases as tests. As a result, it significantly shorted the transfer of typings.</li><li>The proposed tool (dts-gen) didn’t work as expected in all scenarios. For example, it crashes if you are trying to create a definition for one of the sub-packages (i.e., webpack-blocks library). I found out that you still could create a definition for the whole package and then break them into separate sub-packages.</li><li>Default step-by-step instruction in DefinitelyTyped provided some help, but many topics were unclear at first. I had to do some research on those. During this struggle, I realized that my efforts should not vanish in vain. So I decided to write my instruction and post it in <a href="https://levelup.gitconnected.com/publishing-typings-to-definitelytyped-d4e0777e40f5?source=friends_link&amp;sk=da081c9b7e83a1d380e88c09e6595af9">another article</a>.</li></ol><h3>Results</h3><p>It took 1,5 months to move them to DefinitelyTyped repo. During this process, I created ~40 pull-requests. Some reviewers didn’t answer in two weeks, so I had to recreate pull-request and go through the same process again.</p><p>To fully understand the value of this project, I decided to gather data on downloads. I wrote down the number of downloads per package that I contributed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1007/1*p2HP4_QSXpVRwPtVdXugFA.png" /><figcaption>The graphic shows the distribution of downloads per package over time.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1008/1*c0cyjliOI0JBGRQEiIVJIw.png" /><figcaption>This graph shows the total amount of downloads per week.</figcaption></figure><p>Some contributed definitions provided significant value to the community. If you check the above graph, you see that jsdom-global and isomorphic-form-data packages generated half of the downloads.<br>Of course, packages like React or Webpack produce 9 million downloads per week. Still, I think that my work helps some folks to do their work faster.</p><h4>Value for the company</h4><p>We got significant feedback from reviewers and other contributors. Some of the packages provided by us were updated recently by other users.<br>We deepened and expanded definitions of libraries. From the start, we didn’t adequately describe our definitions. But due to strict rules, I had to expand them.</p><p>We created test cases for definitions. It protects our use cases in the future.<br>It also simplified version control for definitions. It matches easier to keep things clear if we use the exact version of the NPM package than a link to a folder in the local repository.</p><h4>Personal value</h4><p>During this project, I deepened my knowledge in advanced topics of Typescript and package types.</p><p>I also want to note the rising feeling of self-worth that you write a code that is used not only by you and your colleagues, but the satisfaction that you participated in the popularization of Typescript.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=20b45141554" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/benefits-of-transferring-types-to-definitelytyped-20b45141554">Benefits of transferring TypeScript types to DefinitelyTyped</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Publishing Typings to DefinitelyTyped]]></title>
            <link>https://levelup.gitconnected.com/publishing-typings-to-definitelytyped-d4e0777e40f5?source=rss-63caf087450b------2</link>
            <guid isPermaLink="false">https://medium.com/p/d4e0777e40f5</guid>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[front-end-development]]></category>
            <category><![CDATA[github]]></category>
            <category><![CDATA[typescript]]></category>
            <category><![CDATA[javascript]]></category>
            <dc:creator><![CDATA[Max Boguslavskiy]]></dc:creator>
            <pubDate>Fri, 01 Nov 2019 10:16:01 GMT</pubDate>
            <atom:updated>2019-11-10T14:47:19.576Z</atom:updated>
            <content:encoded><![CDATA[<h3>Intro</h3><p>Typescript is a great addition to JavaScript. Unfortunately, not every JS package has definitions for TypeScript. Of course, you could create your custom definition in your project, but I believe that sharing these definitions makes things easier for the community:</p><ul><li>A large community maintains shared typings</li><li>The review process could provide valuable insights</li><li>The definition helps spread the Typescript language</li></ul><h3>Publishing</h3><p>Before you start, I assume that you encountered one of the following situations:</p><ul><li>DefTyped types and libraries don’t contain the @type that you want to add</li><li>Existing typings should be updated</li><li>The package provides its definition, and you should remove the DefinitelyTyped definition</li></ul><h3>Preparation</h3><ol><li>Fork the DefinitelyTyped repository</li><li>Clone it locally and create a branch for your change</li></ol><h3>Implementation of the package definition</h3><p>I split the process into three types of activities: add, update, or remove.</p><h4><strong>Add a new package</strong></h4><ol><li>Create a folder for library that you want to type in types directory — types/&lt;name_of_package&gt;.</li><li>Prepare the folder structure. Minimal requirements: index.d.ts,&lt;name_of_package&gt;-tests.ts, tsconfig.json, tslint.json .You could copy them from another package, but don’t forget to update the info!</li><li>Investigate which type of package you are trying to describe. You could research types of packages here: <a href="https://www.typescriptlang.org/docs/handbook/declaration-files/templates.html">https://www.typescriptlang.org/docs/handbook/declaration-files/templates.html</a></li><li>Start with writing tests. They provide some protection for your type of usage. It is quicker to describe typings this way and check that everything is fine. Requirements for the file name: &lt;name of package&gt;-tests.ts. If you want to provide several test files, please move them to the tests folder.</li><li>Write the definition file (index.d.ts). You could create draft typing with dts-gen.</li><li>Add a description to your typing. Provide the name and version of the packaging for your typing. It should match with the package on npmjs.com. Link to the package on Github. List the contributors. Optionally, you could provide a Typescript version.</li><li>Write the tslint file. By default, this file contains the only link to the parent file with linting rules dtslint/dt.json. You shouldn’t override default linting rules without reason.</li><li>Write tsconfig.json. Include * ts files to the tsconfig “files” section.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/721/1*0de4AsaBJ2nHE8T2caxPLA.png" /><figcaption>Example of index.d.ts file.</figcaption></figure><h4><strong>Edit an existing package</strong></h4><p><em>Non-breaking changes</em></p><ol><li>Update the definition. Check that tests cover your package usage. Add new tests if your updates provide new arguments or methods.</li><li>Update definitions. Add yourself as a contributor. Update the version in the description if it’s required. The minor package version was updated. Skip the “patch” version.</li></ol><p><em>Breaking changes</em></p><ol><li>Read the instruction for breaking changes: <a href="https://github.com/DefinitelyTyped/DefinitelyTyped#if-a-library-is-updated-to-a-new-major-version-with-breaking-changes-how-should-i-update-its-type-declaration-package">(link)</a></li><li>Provide tests for a new version.</li><li>Store the old version in the folder. The name for folder must be: v&lt;its major version number&gt;.</li><li>Add this folder in paths in tsconfig.json.</li><li>Update the definitions. Add yourself as a contributor. Update the version in the description if it’s required. The minor package version was updated. Skip “patch” version.</li></ol><h4><strong>Removal of and existing package:</strong></h4><ol><li>Delete typing from ./types</li><li>Add the object to notNeededPackages.json in the root directory.</li><li>Template <a href="https://github.com/sindresorhus/create-html-element">(link)</a>:</li></ol><pre>{ <br>  “libraryName”: “create-html-element”, <br>  “typingsPackageName”: “create-html-element”, <br>  “sourceRepoURL”: “sindresorhus/create-html-element”, <br>  “asOfVersion”: “2.1.0” <br>}</pre><h3>Submit changes to DefinitelyTyped</h3><ol><li>Commit changes</li><li>Run npm test to check that everything went fine.</li><li>Check linting with npm run lint &lt;name of package&gt;</li><li>Fix errors</li><li>Push changes</li><li>Go to Github via the link in push commit</li><li>Create a pull request</li><li>Validate via the checklist that everything is ok</li><li>Check general checkbox and mark checkboxes in the appropriate section</li><li>Wait for a merge or comments from the reviewer</li></ol><h3>Recommendations:</h3><ol><li>@packages. DTS-gen cannot create proper typings for sub-packages (i.e. @webpack-block/assets). It’s a known bug. As a solution, create typings for parent package and copy everything required to typings for subpackage.</li><li>Re-use other typings if you could.</li><li>For global typings, you should avoid “import” in tests. This way, you could check that everything works fine. Check guide: <a href="https://github.com/DefinitelyTyped/DefinitelyTyped#how-do-i-write-definitions-for-packages-that-can-be-used-globally-and-as-a-module">(link)</a></li><li>Github provides excellent tools to deal with merging. You could merge simple conflicts via github.com.</li></ol><h3>Guides</h3><ol><li>Best practices | DefinitelyTyped <a href="http://definitelytyped.org/guides/best-practices.html">(link)</a></li><li>Contribution guide | DefinitelyTyped <a href="http://definitelytyped.org/guides/contributing.html">(link)</a></li><li>Introduction · TypeScript <a href="https://www.typescriptlang.org/docs/handbook/declaration-files/introduction.html">(link)</a></li><li>DefinitelyTyped/DefinitelyTyped <a href="https://github.com/DefinitelyTyped/DefinitelyTyped">(link)</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d4e0777e40f5" width="1" height="1" alt=""><hr><p><a href="https://levelup.gitconnected.com/publishing-typings-to-definitelytyped-d4e0777e40f5">Publishing Typings to DefinitelyTyped</a> was originally published in <a href="https://levelup.gitconnected.com/">Level Up Coding</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>