<?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 Nagesh Kumar on Medium]]></title>
        <description><![CDATA[Stories by Nagesh Kumar on Medium]]></description>
        <link>https://medium.com/@nagesh.monu?source=rss-e93cf777c783------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*pPkh7FonXV4yl8j1-HJlyw.jpeg</url>
            <title>Stories by Nagesh Kumar on Medium</title>
            <link>https://medium.com/@nagesh.monu?source=rss-e93cf777c783------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 03:20:53 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@nagesh.monu/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[P<x> Manager & what’s in a name?]]></title>
            <link>https://medium.com/@nagesh.monu/p-x-manager-whats-in-a-name-576c428388d6?source=rss-e93cf777c783------2</link>
            <guid isPermaLink="false">https://medium.com/p/576c428388d6</guid>
            <category><![CDATA[sdlc]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[project-manager]]></category>
            <category><![CDATA[product-manager]]></category>
            <dc:creator><![CDATA[Nagesh Kumar]]></dc:creator>
            <pubDate>Fri, 20 Jul 2018 13:39:17 GMT</pubDate>
            <atom:updated>2018-07-20T13:46:23.441Z</atom:updated>
            <content:encoded><![CDATA[<p><strong><em>P</em></strong><em>roduct , </em><strong><em>P</em></strong><em>rogram &amp; </em><strong><em>P</em></strong><em>roject</em></p><p>After <strong>P</strong> for <strong><em>Pyramid</em></strong> aka <strong><em>Test Pyramid </em></strong>I thought of understanding role of <strong>P</strong> in <strong><em>P</em></strong>roduct, <strong><em>P</em></strong>rogram &amp; <strong><em>P</em></strong>roject Manager.</p><p>Being in an analogy world, I landed into an analogy of Computer Science, Information Science &amp; Information Technology. I was about to fall into an infinite loop of analogy. Thanks to Agile, it pushed extra blood to brain and I decided to follow incremental approach to understand this <strong><em>P</em></strong>.</p><h4>Iteration 0 — Planning is required prior to work start.</h4><ul><li>Understanding the SDLC.</li><li>Looking at too many <strong><em>P</em></strong>s, I decided to keep P for Program Manager out of scope.</li><li>Responsibilities of The Project and Product Manager.</li><li>Responsibilities in Agile software development.</li></ul><h4>Iteration 1 — What is SDLC?</h4><p>Software Development Life Cycle (SDLC) is a process used by the software industry to design, develop and test high quality softwares. The SDLC aims to produce a high-quality software that meets customer expectations, reaches completion within times and cost estimates.</p><p>The SDLC process involves several distinct stages:</p><ul><li>Planning and Requirement Analysis.</li><li>Defining Requirements.</li><li>Designing the Product Architecture.</li><li>Building or Developing the Product.</li><li>Testing the Product.</li><li>Deployment in the Market and Maintenance.</li></ul><p>There are various software development life cycle models defined and designed which are followed during the software development process.</p><ul><li>Waterfall.</li><li>Iterative.</li><li>Spiral</li><li>V-Model</li><li>Agile</li></ul><h4>Iteration 2— Responsibilities of the Product Manager</h4><ul><li>Responsible for leading the team to deliver the end product.</li><li>Responsible for strategy, the product roadmap, as well as the feature definition of the product.</li><li>Responsible for creating product life cycle for delivering the product.</li><li>Responsible for leading the product team.</li><li>Responsible for working directly with customers.</li></ul><h4>Iteration 3— Responsibilities of the Project Manager</h4><ul><li>Responsible for delivering a successful project.</li><li>Responsible for leading the team.</li><li>Responsible for strategy and the process flow of the project.</li><li>Responsible for the project plan as well as the requirement for the project.</li><li>Responsible for creating project life cycle.</li></ul><h4>Iteration 4— <strong>Responsibilities of a(n) &lt;Owner&gt; (what’s in a name?)</strong></h4><p>Agile is based on the adaptive software development methods, whereas the traditional SDLC models like the waterfall model is based on a predictive approach.</p><p>Predictive teams in the traditional SDLC models usually work with detailed planning and have a complete forecast of the exact role, tasks and features to be delivered in the next few months or during the product life cycle.</p><p>Now a days Agile has become a major driving force behind software development in many organisations and the Project Manager, Process Manager and Product Manager are trying to find their role boundaries.</p><p><strong>Given</strong></p><p>Agile methods are being widely accepted in the software world.</p><p><strong>When</strong></p><p>You don’t want to get into the debate of Project Manager or Product Manager.</p><p><strong>Then</strong></p><p>You need to look at responsibilities of a(n) &lt;Owner&gt; in Agile Methods.</p><p>Me being integral part of Agile software development for more than a decade, I thought off listing down the responsibilities.</p><ul><li>Responsible for project vision — Why? What problem does it solve? Who?</li><li>Responsible for understanding stakeholders needs, capturing new ideas.</li><li>Responsible for making sure the team owns the delivery.</li><li>Responsible for managing product backlog.</li><li>Responsible for deciding new feature will get build or not.</li><li>Responsible for sequencing the feature.</li><li>Responsible for communication between stakeholders and team.</li><li>Responsible for prioritization keeping value and size of a new feature.</li><li>Responsible for understanding value from stakeholder and size from the team and with time improving on assessing the size and value.</li><li>Responsible for backlog grooming with team and sometime involving stakeholders.</li><li>Responsible for identifying and balancing short term vs long term thinking for project.</li><li>Responsible for finding the balance <strong><em>Building the right thing</em></strong>, <strong><em>Build the thing right</em></strong> and <strong><em>Build it fast</em></strong>.</li><li>Responsible for realistic, expectation management with stakeholder and team.</li><li>Responsible for identifying team velocity or combined velocity of all the teams.</li></ul><p>In Agile Software Development <strong>&lt;Owner&gt;</strong> roles are greatly aligned on types of Assignment being executed but one need to be aware of <strong>MUSTs</strong>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=576c428388d6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[RESTful Patterns & Principles]]></title>
            <link>https://medium.com/@nagesh.monu/restful-patterns-principles-69836c2df5d3?source=rss-e93cf777c783------2</link>
            <guid isPermaLink="false">https://medium.com/p/69836c2df5d3</guid>
            <category><![CDATA[api]]></category>
            <category><![CDATA[thoughtworks]]></category>
            <category><![CDATA[json-api]]></category>
            <category><![CDATA[restful]]></category>
            <dc:creator><![CDATA[Nagesh Kumar]]></dc:creator>
            <pubDate>Tue, 10 Jul 2018 06:36:04 GMT</pubDate>
            <atom:updated>2018-07-11T11:43:51.030Z</atom:updated>
            <content:encoded><![CDATA[<p>As RESTful web services don’t follow a prescribed standard except for HTTP, it’s important to build your RESTful API in accordance with industry best practices. The following describes patterns and principles that commonly apply to RESTful design concerns. While they are not written in stone, the following should be followed unless there is a valid reason to make an exception.</p><h3><strong>JSON Design</strong></h3><p>The principle concern with designing the contract is on <strong><em>forward compatibility</em></strong><em>, </em>or designing with an eye towards making future changes additive rather than breaking. An additive change would be adding fields to the existing contract. A breaking change would be removing or renaming fields, or changing the hierarchy. Guessing the ways the API will change is difficult; however, there are a few general patterns to think about:</p><h4><strong><em>Always return JSON objects, not arrays.</em></strong></h4><p>If you return an array, you will never be able to add metadata (e.g. paging) without a breaking change, because you will first have to change the array to an object.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/702/1*akD-oOCZsQZoaxduVMi-fQ.png" /></figure><h4><strong><em>Prefer objects over literals for mutually exclusive options.</em></strong></h4><p>This is especially true when there are multiple fields in each mutually exclusive option. In the air ticket purchase example, some of the fields only apply if you purchase with points, some only apply if you purchase with currency, and some always apply.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TvjA0_GyEPY_GLwCKaJo-w.png" /></figure><h4><strong><em>Prefer arrays over numbered values</em></strong></h4><p>In the air ticket purchase example, for connecting flights booking preferring <strong><em>destination-airport1</em></strong> and <strong><em>destination-airport2</em></strong> for connections. That’s harder to parse as a client, and not as flexible without some tricky and non-idiomatic code.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2P3eh1RUjeth5_Xb2xH_IA.png" /></figure><h4>Prefer hypermedia over flags to expose possible state transitions</h4><p>There are a number of scenarios where the consumer needs to know if a state transition is allowed or not In the air ticket purchase example, to enable or disable a Checkin button on a UI. Using a flag (checkinEligible) or hypermedia both solve the problem.</p><p>Hypermedia has the added advantage of helping client developers know the URL template to integrate against. This is true even if it’s unlikely that they programmatically follow the link. The hypermedia effectively acts as inline documentation to help the client understand consumption patterns of the API.</p><p>However, we’re not aiming for complete programmatic <a href="http://en.wikipedia.org/wiki/HATEOAS">HATEOAS</a> implementation. Prefer relationship keys that give an indication of the state transition (like ‘checkin’ below), but prefer to keep the body and query parameters off of the URL. They are properly part of the media type.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-VUD3CSYmg3XNHUwrlfe-g.png" /></figure><p>There are of course instances where flags are superior to hypermedia.</p><h4><strong><em>Use _ prefix for metadata</em></strong></h4><p>Without some convention, it can be tricky to separate resource data from metadata. Hypermedia is an obvious example of metadata, and we use the _links field to denote it. Air ticket change is another example where we need metadata. We require a name of any passenger on the reservation as a simple sanity check to validate the request, and we require information whether the fund balance goes to travel funds or as a refund. Neither of those are updatable fields on the resource, so we want to wrap them in some metadata container.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3PhP5vgD9Gz9zokiqJXRPA.png" /></figure><p>There is no way to discern metadata attributes from resource attributes.</p><h4><strong><em>Never use query parameters for POSTs and PUTs. Never use body parameters for GETs and DELETEs</em></strong></h4><p>In general, it’s awkward to have to put request parameters in both the querystring and the body, as each is intended to solve the same problem of request parameterization on its own.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8x4_3H_joMEEwNFGI92ILg.png" /></figure><h4><strong><em>Use POST for side-effecting updates</em></strong></h4><p>In the RESTful style that maps CRUD operations to the uniform interface, PUT is the standard verb for updates (a useful alternative to the CRUD style is reifying immutable domain intents as mentioned above). However, the underlying assumption is that, in the event of a communication failure that prevents the client from getting the server response, the client should always be able to retry a PUT. If the implementation has side-effects (e.g. debiting a credit card), it is better to tunnel the call through a POST than send the update through a PUT.</p><h4>Avoid floating point numbers in JSON</h4><p>JSON number types support floating point, but the precision will be determined by the consuming platform. Use ints as much as possible. We do this for currency by representing money in cents.</p><h4>Use strings for numbers when math doesn’t make sense</h4><p>When the consumer may use numbers for math, such as money, use ints. Math doesn’t always make sense — for example, account numbers and phone numbers. Use strings for those fields.</p><h4>Always use ISO datetimes</h4><p>Simplify date validation by using ISO format for all dates.</p><h3>URI Design</h3><p>In REST, the URL path (e.g. without the querystring) acts as an id for the resource. As such, all resourceful URL paths should avoid the use of verbs.</p><p>Oversimplifying a bit, there are <strong><em>five types</em></strong> of endpoints:</p><p><strong>1. List resources (e.g. /reservations)</strong></p><ul><li>These represent a collection of resources.</li><li>They should be plural nouns with no identifier in the URL path.</li><li>List resources can be filtered, sorted, etc. using query parameters. Any search that can return more than 1 result should be represented as query parameters on a list resource. <strong>Example</strong>: /reservations?credit-card=123456789.</li></ul><p><strong>2. Singular resources (e.g. /reservations/record-locator/ABC123)</strong></p><ul><li>These represent a single instance of a resource in a collection.</li><li>They should be completely and uniquely identified on the URL path, and should use a <em>path-indicator</em> between the list resource and the identifier. The path-indicator allows flexibility for using different mechanisms to identify the resource. In the example above, <em>record-locator</em> is the path-indicator.</li><li>The key can be a made up composite key where we have no valid shadow key. Product-id is an example.</li></ul><p><strong>3. Nested resources (e.g. /reservations/record-locator/ABC123/mobile-boarding-passes)</strong></p><ul><li>These represent separately addressable resources under a parent singular resource. They are occasionally useful to extract out when you want an independent workflow around the child resource (e.g. checking in).</li><li>Note that we have not followed this principle to date in the code, adding an additional layer (e.g. /reservations/record-locator/ABC123/<em>operation-infos</em>/mobile-boarding-pass). The <em>operation-infos</em> is unnecessary, since the domain of the URL scopes the request to air-travel. We also did not nest the resource under /reservations, which we should to keep it consistent with other reservation resources.</li><li>Try to avoid nesting more than 2 levels.</li></ul><p><strong>4. Reified resources</strong> <strong>(e.g. /reservations/record-locator/ABC123/changeReservationRequest)</strong></p><ul><li><em>Reification </em>is the act of naming an abstract concept, which is sometimes useful in RESTful modeling. Reified resources generally represent the intent of a change, as opposed to a simple CRUD operation. As such, they tend to avoid PUTs in favor of immutable resources modeling some user action.</li></ul><p><strong>5. Non-resourceful endpoints (e.g. /reservations/search)</strong></p><ul><li>There may be considerations that force us into an intentionally non-resourceful endpoint. For example, we decided to support searching for reservations by credit cards as a POST instead of a GET because we want to keep the credit card number off of the URL. In this case, we are clearly and intentionally breaking the Uniform Interface architectural constraint of REST. To clearly indicate this as intentional, we should put a verb on the path to label this as RPC instead of RESTful.</li></ul><h3>Status Codes</h3><p><strong>200 Ok</strong></p><ul><li>In some contexts, this can be an empty result from a search on a list resource. That’s important to note, because some discussions lean towards returning a 400 when the resultset is empty. Think of it this way — if you somehow manage to enter a Google search for something that Google has never heard of, the correct response from Google would be to return an empty list of search results back, not to send back indicated that you searched for an invalid term.</li></ul><p><strong>201 Created</strong></p><ul><li>Should be used as the result of POSTs to list resources, as it indicates a new resource has been created. Should contain a Location header providing the URL for the new resource.</li><li>Prefer to send back the same response body of the newly formed resource as a GET to the Location header would provide, to save the client a call.</li></ul><p><strong>202 Accepted</strong></p><ul><li>Used for asynchronous operations, and clearly indicates eventual consistency.</li></ul><p><strong>204 No Content</strong></p><ul><li>Like a 200, but with no response body.</li></ul><p><strong>304 Not Modified</strong></p><ul><li>Used in conditional GET operations as part of a caching strategy. The client can send either an ETag or If-Modified-Since header representing the last time they saw the resource, and the server can avoid sending the response body if the resource has not changed.</li></ul><p><strong>400 Bad Request</strong></p><ul><li>Indicates a client mistake, either a programming mistake or an end user mistake.</li></ul><p><strong>401 Unauthorized</strong></p><ul><li>Indicates an unauthenticated request.</li></ul><p><strong>403 Forbidden</strong></p><ul><li>The server is refusing to fulfill the request.</li></ul><p><strong>404 Not Found</strong></p><ul><li>The resource does not exist.</li><li>Indicates the server does not wish to reveal exactly why the request has been refused, or when no other response is applicable.</li></ul><p><strong>409 Conflict</strong></p><ul><li>If we use optimistic concurrency (e.g. a version or timestamp field), this could represent a concurrency conflict.</li></ul><p><strong>500 Internal Server Error</strong></p><p><strong>503 Service Unavailable</strong></p><ul><li>Used to indicate a downstream dependency failed.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=69836c2df5d3" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>