<?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 Rishi Raj Jain on Medium]]></title>
        <description><![CDATA[Stories by Rishi Raj Jain on Medium]]></description>
        <link>https://medium.com/@rishi-raj-jain?source=rss-cf48f7f14d64------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*LaMgMWb4pcTf7hzovRuLfw.jpeg</url>
            <title>Stories by Rishi Raj Jain on Medium</title>
            <link>https://medium.com/@rishi-raj-jain?source=rss-cf48f7f14d64------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 16:11:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rishi-raj-jain/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[The Best Postgres Extensions (And How to Test Them on Production Data Safely)]]></title>
            <link>https://medium.com/@rishi-raj-jain/the-best-postgres-extensions-and-how-to-test-them-on-production-data-safely-64ecf622f0c5?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/64ecf622f0c5</guid>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Thu, 14 May 2026 15:22:11 GMT</pubDate>
            <atom:updated>2026-05-14T15:22:11.622Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LbKXzkRzE6_hFbd2wah1QA.png" /></figure><p>Postgres extensions can replace whole services in your stack such as vector databases, job queues, search engines. The hard part isn’t choosing them. It’s enabling them on production without breaking it.</p><p>You don’t want to install a random C-library extension on your production database and watch it crash. You don’t know if the version your cloud provider supports matches what you tested locally. And you definitely can’t tell if pgvector&#39;s similarity search will actually work on <em>your</em> data until you try it, but trying it in production isn&#39;t an option.</p><p>This is where teams would get stuck. They know powerful Postgres extensions exist, they just can’t safely validate them before fully committing. This guide is about that validation step i.e. how to test any Postgres extension against your real production data without putting production at risk.</p><blockquote><em>If you’re still deciding which extensions are worth adopting, the post </em><a href="https://xata.io/blog/is-postgres-really-enough-in-2026"><em>Is Postgres really enough in 2026?</em></a><em> covers the case for replacing Redis, Elasticsearch, and MongoDB with Postgres primitives. This article picks up where that one stops — given you’ve decided to enable an extension, how do you ship it safely?</em></blockquote><h3>Why “works on my machine” fails for Postgres extensions</h3><p>You spin up a local Docker container with pgvector v0.5.0. Everything runs smoothly. You deploy, and your cloud provider&#39;s managed Postgres is on v0.4.0. The index syntax had changed between versions, and your queries fail on the first request in production.</p><p>This is not hypothetical. <strong>Extension version mismatches cause outages</strong>, and you won’t catch them in staging if your staging database is empty or seeded with synthetic data.</p><p>The problem is deeper than versions. Most extensions have behavior that only shows up under real data:</p><p>Extension What only real data tells you pgvector Whether your similarity threshold returns relevant results, and whether the index fits in memory pg_trgm Whether your similarity threshold catches typos without flooding results with noise pg_cron Whether your scheduled job actually completes before the next run kicks off postgis Whether your spatial queries are sub-second on real-world geometries pg_repack Whether bloat reclamation finishes in your maintenance window pg_stat_statements Which queries are actually expensive under production traffic shapes</p><p>Synthetic data won’t surface any of this. Lorem Ipsum doesn’t have typos. Generated embeddings don’t have semantic structure. Faked GPS points don’t cluster the way real users do. You need production-shaped data to know whether an extension will hold up, and you need to test it somewhere that isn’t production.</p><h3>How to safely test pgvector on production-like data</h3><p>If you may only test one extension against production data, make it pgvector. Every other extension on the list has some meaningful local test - citext either works or it doesn&#39;t, pg_cron&#39;s scheduler is deterministic, and uuid-ossp is just a function call. pgvector is the one where everything that matters is invisible without real data.</p><p>Consider what you have to decide before shipping a vector search feature:</p><ol><li><strong>Index type and parameters</strong>: ivfflat vs. hnsw. lists for ivfflat or m and ef_construction for hnsw. The wrong choice can mean a 10× difference in query latency or a 10-point drop in recall.</li><li><strong>Similarity threshold</strong>: What cosine distance counts as “relevant”? On random vectors this is meaningless. On your actual embeddings, the answer changes per content type.</li><li><strong>Memory footprint</strong>: A 1536-dimensional hnsw index on a million rows is roughly 6–8 GB. Will it fit in shared buffers on your production instance, or will every query hit disk?</li><li><strong>Recall vs. latency</strong>: ivfflat lets you trade probes (more probes = higher recall, slower queries). The right number depends entirely on how your embeddings cluster.</li></ol><p>None of these can be answered with INSERT INTO documents VALUES (random_vector(1536)). Random vectors are uniformly distributed while real embeddings cluster around topics, languages, and document types. The whole point of pgvector is to exploit that clustering. Test on random data and you&#39;re testing the wrong thing.</p><p>Here’s the minimal setup for using pgvector:</p><pre>CREATE EXTENSION IF NOT EXISTS vector;</pre><pre>ALTER TABLE documents ADD COLUMN embedding VECTOR(1536);</pre><pre>-- After backfilling embeddings from your real content...</pre><pre>CREATE INDEX documents_embedding_idx<br>  ON documents USING ivfflat (embedding vector_cosine_ops)<br>  WITH (lists = 100);</pre><pre>SELECT id, content,<br>       1 - (embedding &lt;=&gt; $1) AS similarity<br>FROM documents<br>ORDER BY embedding &lt;=&gt; $1<br>LIMIT 10;</pre><p>The questions you actually need to answer:</p><ul><li>Do the top 10 results look relevant to a human reviewer for at least 20 representative queries?</li><li>What’s the p95 latency at production-scale row counts?</li><li>What’s the index size and does it fit in shared_buffers?</li><li>If you change lists from 100 to 1000 (or switch to hnsw), how do recall and latency move?</li></ul><p>You cannot answer any of those without your actual documents and your actual embedding model, which brings us to the following workflow.</p><h3>Test extensions on a branch of production data</h3><p>This workflow applies to any extension you want to safely evaluate before deploying to production:</p><ol><li>Branch production into an isolated environment that holds production-shaped data.</li><li>Enable the extension in that branch only.</li><li>Run the queries you’d run in production. Measure, tune and iterate.</li><li>Once validated, promote the same change to production with confidence.</li></ol><p><a href="https://xata.io/docs/core-concepts/branching">Xata’s database branching</a> is built for exactly this. Branches are copy-on-write, so creating one is instant and cheap. The branch shares storage with its parent until it diverges and <a href="https://xata.io/postgres-data-masking">built-in anonymization</a> means you can use the branch to test without exposing real customer data.</p><h3>1. Branch production</h3><pre>xata branch create test-pgvector --parent-branch main</pre><p>This creates <a href="https://xata.io/docs/core-concepts/branching#copy-on-write">a copy-on-write branch</a> of your production database. You now have a database with the same schema, the same row counts, the same data distribution, and the same query patterns as production but available with a different connection string.</p><h3>2. Enable the extension</h3><pre># Connect to the branch<br>psql &quot;$(xata branch url --branch test-pgvector)&quot;</pre><pre>-- Enable the extension<br>CREATE EXTENSION IF NOT EXISTS vector;</pre><p>The extension is enabled in the branch only. Production is untouched. If the extension fails to install, fails to load, or conflicts with another extension, you find out here itself.</p><h3>3. Test for real</h3><p>This is the step that synthetic data can’t replicate. Now, you can run your actual queries against the branch:</p><pre>ALTER TABLE documents ADD COLUMN embedding VECTOR(1536);</pre><pre>-- Backfill embeddings from your application or a migration script<br>-- (this also tells you how long the backfill takes on real volume)</pre><pre>CREATE INDEX documents_embedding_idx<br>  ON documents USING ivfflat (embedding vector_cosine_ops)<br>  WITH (lists = 100);</pre><pre>-- Run the queries that production will run<br>EXPLAIN (ANALYZE, BUFFERS)<br>SELECT id, content,<br>       1 - (embedding &lt;=&gt; $1) AS similarity<br>FROM documents<br>ORDER BY embedding &lt;=&gt; $1<br>LIMIT 10;</pre><p>Things you can now check that you couldn’t before:</p><ul><li><strong>Index build time on real volume</strong>: If it takes 4 hours on the branch, you can schedule the production rollout accordingly.</li><li><strong>Query performance and latency</strong>: You can expect it to closely match to what it would show up as in real production data scale.</li><li><strong>Result relevance</strong>: Sample 20 production queries and eyeball the top 10 results. If the answers look wrong, you’ve caught a problem before pushing the extension to production.</li></ul><p>Once you’ve validated the extension in your branch and have documented any changes needed for your specific workflow, you can confidently apply the same steps to your production branch, knowing your rollout will go smoothly and safely!</p><h3>A pre-flight checklist for any extension</h3><p>Before merging an extension change to production, you should be able to confirm all of the following:</p><ul><li>[ ] The extension version available on my managed Postgres matches the version I tested with.</li><li>[ ] CREATE EXTENSION succeeded on a branch with production-shaped data.</li><li>[ ] The extension’s queries return correct, relevant results on real samples (verified by a human, not a unit test).</li><li>[ ] Any indexes the extension creates fit within available memory, or I’ve planned for the disk-backed path.</li><li>[ ] Backfill / migration steps complete within an acceptable window at production volume.</li><li>[ ] Rollback is documented (DROP EXTENSION, column drops, index drops).</li><li>[ ] Monitoring is in place for the queries the extension powers.</li></ul><h3>Conclusion</h3><p>Postgres is not just a relational database. It’s a platform that can replace significant chunks of your infrastructure stack when configured correctly. The extensions exist but the challenge is testing them safely before committing to production.</p><p>To simplify your stack and reduce costs, first see if a Postgres extension can handle tasks you’d normally offload to another service, test the extension safely on a Xata branch with real data, then roll it out to production once validated so that you deploy with confidence, knowing exactly how it would work without breaking production.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=64ecf622f0c5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Build an AI Tutor for Open Source Repositories with DeepWiki]]></title>
            <link>https://medium.com/@rishi-raj-jain/build-an-ai-tutor-for-open-source-repositories-with-deepwiki-cd66d6012099?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/cd66d6012099</guid>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Tue, 05 May 2026 05:03:10 GMT</pubDate>
            <atom:updated>2026-05-05T05:03:10.639Z</atom:updated>
            <content:encoded><![CDATA[<p>During team calls when someone will ask what a repository does, how an API works, or where two modules connect, the group then pauses while someone searches docs, GitHub issues, and old messages. That context switching makes discussions slower and it can leave newer contributors behind because they do not know where to look.</p><p>This tutorial will help you add an AI tutor to the call so the conversation keeps moving. The tutor listens to what people are asking, looks up the answer in DeepWiki, and replies in simple terms. You can ask quick questions, follow up with more detail, and get an explanation that teaches you the desired concepts. The goal isn’t to replace reading the code, but to save time lost switching context and fill in knowledge gaps so that the team can spend more time on real work instead of waiting for a teammate’s explanation.</p><p>We’ll build this tutor using <a href="https://visionagents.ai/">Vision Agents</a> for the agent framework, <a href="https://getstream.io/">Stream</a> for WebRTC audio and video transport, <a href="https://platform.openai.com/audio/realtime">OpenAI Realtime</a> for speech in and speech out, and <a href="https://anam.ai/">Anam</a> for a visible avatar on the call. <a href="https://deepwiki.com/">DeepWiki</a> is connected through its hosted MCP server, which lets the model use DeepWiki tools during the conversation without authentication. The result is a small piece of code that creates a helpful tutoring presence in the meeting, so people can ask about the codebase without breaking the flow of the discussion.</p><h2>Demo</h2><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F9wREhQu2tX0%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D9wREhQu2tX0&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F9wREhQu2tX0%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/580f906bdb0834d542d82aab8cbbed07/href">https://medium.com/media/580f906bdb0834d542d82aab8cbbed07/href</a></iframe><h2>Prerequisites</h2><p>You will need the following to get going with the implementation:</p><ul><li><a href="https://www.python.org/downloads/">Python 3.12</a> or later (matches typical vision-agents setups)</li><li><a href="https://docs.astral.sh/uv/">uv</a> Python package manager</li><li>A <a href="https://beta.dashboard.getstream.io/">Stream</a> account</li><li>An <a href="https://platform.openai.com/">OpenAI</a> account with access to Realtime-capable models (this walkthrough uses gpt-realtime-1.5)</li><li>An <a href="https://anam.ai/">Anam</a> account</li></ul><h2>Create a Stream Application</h2><ul><li>Navigate to the <a href="https://beta.dashboard.getstream.io/">Stream dashboard</a>.</li><li>Select <strong>+ Create an App</strong>.</li><li>In the dialog, enter a name for your application and choose the appropriate <a href="https://getstream.io/blog/chat-edge-infrastructure/">region</a> for your edge-server location(s).</li><li>After creation, locate the <strong>API Key</strong> and <strong>Secret</strong> under <strong>Your Credentials</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZK8dYjyWDjQ-nu5OeJOL1w.png" /></figure><ul><li>Add these credentials to your .env file as follows:</li></ul><pre><br>STREAM_API_KEY=&quot;your-api-key&quot;<br>STREAM_API_SECRET=&quot;your-api-secret&quot;</pre><h2>Configure OpenAI</h2><ul><li>Access the <a href="https://platform.openai.com/settings/organization/api-keys">OpenAI API Key dashboard</a>.</li><li>Click <strong>+ Create new secret key</strong>.</li><li>Add the generated key to your .env file:</li></ul><pre>OPENAI_API_KEY=&quot;your-openai-api-key&quot;</pre><p>Ensure your organization can use the <strong>Realtime API</strong> and the model you pick in code (here, gpt-realtime-1.5).</p><h2>Configure Anam</h2><ul><li>Access the <a href="https://lab.anam.ai/api-keys">Anam API Key dashboard</a>.</li><li>Click <strong>+</strong> to create a new key.</li><li>Add the generated key to your .env file:</li></ul><pre>ANAM_API_KEY=&quot;your-anam-api-key&quot;</pre><ul><li>Open <a href="https://lab.anam.ai/build">Anam Build</a>.</li><li>Click <strong>Avatar</strong>, hover your desired avatar, and click <strong>Copy ID</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JrKDLZYZb6eXYeHPJ9xIzg.png" /></figure><ul><li>Add the ID to your .env file:</li></ul><pre>ANAM_AVATAR_ID=&quot;your-anam-avatar-id&quot;</pre><p>That’s it for configuring your environment variables. Next, let’s move on to scripting the AI tutor.</p><h2>Set up a new Python application</h2><p>In this section, you will learn how to create a new Python application, set up vision agents in it, and install relevant libraries for a quick implementation.</p><p>Let’s get started by creating a new Python project. Open your terminal and run the following commands:</p><pre>mkdir my-ai-tutor &amp;&amp; cd my-ai-tutor<br>uv init &amp;&amp; uv add &quot;vision-agents[anam,getstream,openai,redis]&quot; python-dotenv httpx</pre><p>Now, create a .env file at the root of your project. You are going to add the items we saved from the above sections.</p><p>It should look something like this:</p><pre># .env<br><br>## Stream environment variables<br><br>STREAM_API_KEY=&quot;...&quot;<br><br>STREAM_API_SECRET=&quot;...&quot;<br><br>EXAMPLE_BASE_URL=&quot;https://demo.visionagents.ai&quot;<br><br>## OpenAI (Realtime) environment variable<br><br>OPENAI_API_KEY=&quot;sk-proj-...&quot;<br><br>## Anam environment variables<br><br>ANAM_API_KEY=&quot;...&quot;<br><br>ANAM_AVATAR_ID=&quot;...-...-...-...-...&quot;</pre><p>Then, create a main.py file with the following code:</p><pre>from dotenv import load_dotenv<br>from vision_agents.core import Agent, AgentLauncher, Runner<br>from vision_agents.plugins import openai, getstream<br>from vision_agents.core.edge.types import User<br><br>load_dotenv()<br><br>async def create_agent(**kwargs) -&gt; Agent:<br>    return Agent(<br>        edge=getstream.Edge(),  # GetStream edge network for AV transport<br>        agent_user=User(name=&quot;DeepWiki Tutor&quot;, id=&quot;agent&quot;),  # Agent&#39;s identity<br>        instructions=(<br>            &quot;&quot;&quot;<br>            You are DeepWiki Tutor, an expert teaching agent whose job is to help anyone understand any concept clearly and deeply.<br><br>Your mission:<br>- Teach the user any concept in a structured, beginner-friendly but intellectually serious way.<br>- Explain like a great teacher, not like a textbook.<br>- Prefer clarity, intuition, structure, and examples over jargon.<br>- When jargon is necessary, define it immediately in plain language.<br><br>Core behavior:<br>- Start with a one-sentence plain-English definition of the concept.<br>- Then explain why it matters.<br>- Break the concept into its key parts or first principles.<br>- Show how the parts connect.<br>- Give at least one simple example and, when useful, one real-world example.<br>- Mention common misconceptions or confusion points.<br>- End with a short recap and 2 to 5 natural follow-up questions the user might ask next.<br><br>Teaching style:<br>- Be patient, clear, and encouraging.<br>- Never assume prior knowledge unless the user signals expertise.<br>- Use analogies only when they truly help.<br>- Avoid unnecessary complexity.<br>- If the concept has multiple difficulty levels, present them in layers:<br>  1. Basic intuition<br>  2. Core mechanics<br>  3. Deeper nuance<br>  4. Advanced implications or edge cases<br><br>Adaptation rules:<br>- If the user says “explain simply,” keep it short and intuitive.<br>- If the user says “go deeper,” expand with more rigor, mechanisms, tradeoffs, and edge cases.<br>- If the user is a beginner, avoid dense terminology.<br>- If the user seems advanced, include formal distinct<br>            &quot;&quot;&quot;<br>        ),<br>        llm=openai.Realtime(model=&quot;gpt-realtime-1.5&quot;, voice=&quot;ash&quot;),<br>    )<br><br><br>async def join_call(agent: Agent, call_type: str, call_id: str, **kwargs) -&gt; None:<br>    call = await agent.create_call(call_type, call_id)<br>    # Context manager handles join and clean-up automatically<br>    async with agent.join(call):<br>        await agent.finish()<br><br><br>if __name__ == &quot;__main__&quot;:<br>    # Entrypoint for CLI: launches agent and runs the join_call automatically when invoked from the command line<br>    Runner(AgentLauncher(create_agent=create_agent, join_call=join_call)).cli()</pre><p>The code block above loads the environment variables, then create_agent returns an Agent that connects to Stream using getstream.Edge() and converses in real-time using openai.Realtime with a chosen model and voice. The DeepWiki Tutor name is what people see in the call. join_call creates the call, joins it, and exits cleanly when the call finishes.</p><p>The key to a personalized agent is the instructions text. It tells the model who it is and how to behave, so it consistently teaches instead of sounding like a generic assistant. Because the rules mention structure, definitions, examples, and adapting to the user, the tutor can match beginners or advanced users while staying clear and focused. Strong instructions are what make the same model feel like your tutor rather than a random voice on the call.</p><h2>Bring Your Tutor to Life with an Anam Avatar</h2><p><a href="https://anam.ai/">Anam</a> adds a real time interactive avatar to the agent. The avatar speaks with natural movements and automatic lip sync, which makes a voice agent feel like a present participant on a video call instead of an invisible bot.</p><p>Vision Agents ships an <a href="https://visionagents.ai/integrations/avatars/anam">Anam integration as a processor</a>. A processor in this context is a flexible building block i.e. it can be an avatar like Anam, a computer vision model (like YOLO), or any custom Python code designed to intercept raw WebRTC audio and video streams, perform some manipulation or transformation, and then republish the processed result back into the Stream call. For Anam, the processor takes audio output from the agent, streams it to the Anam service, and then publishes the resulting animated avatar video and audio as part of the group call experience.</p><p>Make the following changes to add Anam avatar to your agent:</p><pre># existing imports<br><br>+ from vision_agents.plugins.anam import AnamAvatarPublisher<br><br>async def create_agent(**kwargs) -&gt; Agent:<br>    ...<br>    agent = Agent(<br>        ...<br>+        processors=[AnamAvatarPublisher()],<br>    )<br><br>    return agent<br><br># rest of the file</pre><p>The code block above integrates Anam avatars into your agent by importing AnamAvatarPublisher and adding it to the agent&#39;s processors list, enabling the agent to produce a real-time animated avatar with expressions in calls.</p><h2>Study Open Source Code with DeepWiki</h2><p><a href="https://deepwiki.com/">DeepWiki</a> is an AI-powered documentation you can talk to for any repository. It helps you understand unfamiliar codebases by letting you explore repo knowledge in a conversational way.</p><p>DeepWiki provides an official remote MCP server that is free for public repositories and does not require authentication. When you connect it, your agent gets DeepWiki tools it can invoke during a conversation, including read_wiki_structure to discover available topics for a GitHub repo, read_wiki_contents to fetch documentation content, and ask_question to ask a repo specific question and get a context grounded answer. The recommended endpoint is the Streamable HTTP MCP route at https://mcp.deepwiki.com/mcp (<a href="https://docs.devin.ai/work-with-devin/deepwiki-mcp">source</a>).</p><p>But what’s MCP? <a href="https://modelcontextprotocol.io/docs/getting-started/intro">MCP (Model Context Protocol)</a> is a standard way for an agent to connect to external tools over a single interface. The big win is that you do not have to hand write tool functions inside your app for each capability. You register an MCP server, the agent learns what tools exist, and the model can call them when it needs grounded information.</p><p>Make the following changes to add the DeepWiki MCP server so the tools are registered with your agent:</p><pre># existing imports<br><br>+ from vision_agents.core.mcp import MCPServerRemote<br><br>async def create_agent(**kwargs) -&gt; Agent:<br>    ...<br>    agent = Agent(<br>        ...<br>+        mcp_servers=[<br>+            MCPServerRemote(<br>+                url=&quot;https://mcp.deepwiki.com/mcp&quot;,<br>+                timeout=10.0,<br>+            )<br>+        ],<br>    )<br><br>    return agent<br><br># rest of the file</pre><p>VisionAgents SDK makes registering a remote MCP server very easy, mostly just a configuration. You declare the remote endpoint URL and optionally include headers for authentication when a server requires it, then the framework connects to the MCP server, discovers the tools it exposes, and automatically registers them so the model can call them without you writing wrapper functions for each tool. (<a href="https://visionagents.ai/guides/mcp-tool-calling#remote-servers">source</a>)</p><h2>How to Run</h2><p>To launch the service locally, use the following command in your terminal:</p><pre>uv run main.py run</pre><p>It will open a demo at <a href="https://demo.visionagents.ai/">https://demo.visionagents.ai</a> and automatically join the session for you. You can then interact with your tutor in real time, as demonstrated in the <a href="https://markdown2medium.vercel.app/#demo">demo</a>.</p><p>The terminal will display the following output prior to the agent joining the call, confirming that the tools have been automatically registered via the remote MCP server:</p><pre>01:31:47.321 | INFO     | [Agent: agent] | 🔌 Connecting to 1 MCP server(s)<br>01:31:47.321 | INFO     | [Agent: agent] |   Connecting to MCP server 1/1: MCPServerRemote<br>01:31:47.322 | INFO     | Connecting to remote MCP server at https://mcp.deepwiki.com/mcp<br>01:31:48.368 | INFO     | Successfully connected to remote MCP server at https://mcp.deepwiki.com/mcp (session: None)<br>01:31:48.368 | INFO     | [Agent: agent] |   ✅ Connected to MCP server 1/1<br>01:31:50.054 | INFO     | [Agent: agent] |   📋 Found 3 tools from MCP server 1<br>01:31:50.054 | INFO     | [Agent: agent] |     ✅ Registered tool: mcp_0_read_wiki_structure<br>01:31:50.054 | INFO     | [Agent: agent] |     ✅ Registered tool: mcp_0_read_wiki_contents<br>01:31:50.054 | INFO     | [Agent: agent] |     ✅ Registered tool: mcp_0_ask_question<br>01:31:50.055 | INFO     | [Agent: agent] | 🤖 Agent joining call: ...-...-...-a933-861192557a57<br>01:31:50.055 | INFO     | Added 3 tools to session config: [&#39;mcp_0_read_wiki_structure&#39;, &#39;mcp_0_read_wiki_contents&#39;, &#39;mcp_0_ask_question&#39;]<br>01:31:51.851 | INFO     | ✅ RTC connection established<br>01:31:52.083 | INFO     | [Agent: agent] | 🤖 Agent joined call: ...-...-...-a933-861192557a57</pre><h2>Ending Thoughts</h2><p>It’s exciting how easy it is to connect real-time speech, a talking avatar, and smart AI that can search and teach real code. In just a few steps with <a href="https://visionagents.ai/">VisionAgents SDK</a>, you can make an agent that chats with you, and finds answers from real projects as you talk. This means you can actually learn from real code as you have a conversation.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cd66d6012099" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Using Stream to Build a Livestream Chat in Next.js]]></title>
            <link>https://medium.com/@rishi-raj-jain/using-stream-to-build-a-livestream-chat-in-next-js-e23ca1fe29ec?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/e23ca1fe29ec</guid>
            <category><![CDATA[react]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[live-streaming]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Sun, 12 May 2024 10:37:18 GMT</pubDate>
            <atom:updated>2024-05-12T10:37:18.174Z</atom:updated>
            <content:encoded><![CDATA[<p>Learn how to build an engaging and interactive live streaming chat app with Next.js, Stream, and Chatscope.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*92rz8ANGRv3ltGkW" /></figure><p>I always wondered how to create the dynamic chat experience of <a href="https://getstream.io/glossary/live-streaming/">livestreams</a>, like those found on YouTube, but with an added convenience of allowing anyone to participate without logging in.</p><p>With Next.js and <a href="https://getstream.io/chat/sdk/react/">Stream</a>, I was able to successfully create that experience. In this tutorial, I cover a straightforward approach to create a livestream chat environment with granular control over user permissions, making real-time interaction more accessible.</p><h3>Demo</h3><p>Before we jump into the technical stuff, let me give you a sneak peek of what you will be building in this tutorial 👇🏻</p><h3>Rishi Raj Jain on Twitter: &quot;Building reactive chat with @getstream_io and @chatscope in @nextjs, styled with @shadcn ✨It&#39;s so cool to see how I can create a user on demand, post a message and disconnect the user in minimal lines of code.Gonna write a guide on this for sure 💪 pic.twitter.com/LrHaC2mL3j / Twitter&quot;</h3><p>Building reactive chat with @getstream_io and @chatscope in @nextjs, styled with @shadcn ✨It&#39;s so cool to see how I can create a user on demand, post a message and disconnect the user in minimal lines of code.Gonna write a guide on this for sure 💪 pic.twitter.com/LrHaC2mL3j</p><h3>Prerequisites</h3><p>You will need the following to get going with the implementation:</p><ul><li><a href="https://nodejs.org/en/blog/announcements/v18-release-announce">Node.js 18</a> or later</li><li>A <a href="https://getstream.io/try-for-free/">Stream</a> account</li><li>A <a href="https://vercel.com/">Vercel</a> Account</li></ul><h3>Setting up a new Stream application</h3><p>In this section, you will learn how to create a Stream application, enable <a href="https://getstream.io/chat/docs/react/authless_users/">authless users</a>, and define user permissions to read and post into a channel. Let’s get started.</p><p>Once you have created a Stream account and are logged in, click on <strong>+ Create an App</strong> to start creating a Stream chat application.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-KANMc4rRV5CD2CU0_5PDg.png" /></figure><p>Enter an application name of your choice. Further, select the <strong>Chat Storage Location</strong> near to your website deployment region/function. I am in India, so I have set it to Mumbai.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pKZh4_ad5mONRJAlS8AWHQ.png" /></figure><p>Once done, you will be taken to the <strong>Chat Overview</strong> screen. Copy the API Key, and save it somewhere safe to be used further in your Next.js application as <strong>NEXT_PUBLIC_STREAM_API_KEY</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6lXztJfC_XkQx7T-MdWqNQ.png" /></figure><p>To allow visitors post chat messages without authorization, scroll down and <strong>enable</strong> the <strong>Disable Auth Checks</strong> toggle button. Click <strong>Submit</strong> to sync the changes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*47_FS702ocJUDH9qEOjrvw.png" /></figure><p>To allow visitors to post a message (as a user in a particular channel), you would need to enable read and post permissions for them. This comes handy to make sure that you are able to configure the roles and permissions a visitor can assume in your chat application.</p><p>Select Role as user and Scope as messaging, and then click the <strong>Edit</strong> button to start configuring the permissions for a visitor.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QFIWxwYKDZwn5t2ZE--1Rg.png" /></figure><p>To allow visitors to post message, search for <strong>create message</strong> and enable the permission to <strong>Create Message </strong>which allows any user in the channel to send a message.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AfoIRy6Yy9gH57LVW7zvlA.png" /></figure><p>To allow visitors to read the chat history, search for <strong>read channel </strong>and enable the <strong>Read Channel</strong> and <strong>Read Channel Members</strong> which allows a visitor to read the messages along with the users who posted them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*drHJjQZLgfPM80_5gZUxBQ.png" /></figure><p>Finally, click <strong>Save</strong> to sync the changes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KyaZ0h086I_lXlED4sFEjw.png" /></figure><p>You are now done with configuring the permissions and scope of what a visitor can do. Now, let’s build the chat user interface in Next.js with Chatscope UI Kit.</p><h3>Setting up a new Next.js application</h3><p>In this section, you will learn how to create a new Next.js application, set up shadcn/uiin it, and figure out requirements of your application to install relevant libraries for a quick implementation.</p><p>Let’s get started by creating a new Next.js project. Open your terminal and run the following command:</p><pre>npx create-next-app@latest my-chat-app</pre><p>When prompted, choose:</p><ul><li>Yes when prompted to use TypeScript.</li><li>No when prompted to use ESLint.</li><li>Yes when prompted to use Tailwind CSS.</li><li>No when prompted to use src/ directory.</li><li>Yes when prompted to use App Router.</li><li>No when prompted to customize the default import alias (@/*).</li></ul><p>Once that is done, move into the project directory and start the app in development mode by executing the following command:</p><pre>cd my-chat-app<br>npm run dev</pre><p>The app should be running on <a href="http://localhost:3000/">localhost:3000</a>.</p><p>Now, create a .env file at the root of your project. You are going to add the items we saved from the above sections.</p><p>It should look something like this:</p><pre># .env<br><br># Stream Environment Variable<br>NEXT_PUBLIC_STREAM_API_KEY=&quot;...&quot;</pre><h4>Integrating shadcn/ui components</h4><p>To quickly prototype the chat user interface, you will set up the shadcn/ui with Next.js. shadcn/ui is a collection of beautifully designed components that you can copy and paste into your applications. Execute the command below to set up shadcn/ui:</p><pre>npx shadcn-ui@latest init</pre><p>You will be asked a few questions to configure a components.json, choose the following:</p><ul><li>Yes when prompted to use TypeScript.</li><li>Default when prompted to select the style to use.</li><li>Slate when prompted to choose the base color.</li><li>yes when prompted to use CSS variables for colors.</li></ul><p>Once that is done, you have set up a CLI that allows us to easily add React components to your Next.js application. Execute the following command to get the button, input, and textarea elements:</p><pre>npx shadcn-ui@latest add button<br>npx shadcn-ui@latest add input<br>npx shadcn-ui@latest add textarea</pre><p>Once that is done, you would now see a <strong>ui</strong> directory inside the app/components directory containing <strong>button.tsx</strong>, <strong>input.tsx</strong>, and <strong>textarea.tsx</strong> files.</p><h4>Installing Chatscope and Stream UI Kit for React</h4><p>Let’s figure out overall actions of what is needed so that you can install the relevant dependencies to make the implementation easier.</p><p>First, as each visitor is treated as an anonymous user, you would need to identify them with a randomly generated name (id). Another is that each message received (in the channel) shall propagate to all the visitors in real time. Finally, you need to make sure that all messages are represented as incoming ones into the channel.</p><p>Execute the following command to install the necessary libraries for satisfying the requirements as above:</p><pre>npm install @chatscope/chat-ui-kit-react<br>npm install stream-chat stream-chat-react<br>npm install unique-username-generator</pre><p>It installs the following libraries:</p><ul><li>unique-username-generator: A package to generate unique usernames.</li><li>stream-chat: Stream’s JavaScript API client.</li><li>stream-chat-react: A library of React hooks (and components) for Stream Chat.</li><li>@chatscope/chat-ui-kit-react: A library of React components for prototyping Chat UI.</li></ul><p>Now, let’s move on to building the reactive chat user interface.</p><h3>Building the Chat User Interface with ChatScope and Stream</h3><p>In this section, you will learn how to build React components that attach visitors to a specific channel, render the chat history, and allow them to post messages.</p><p>First, let’s build a React component that renders a list of chat messages dynamically. You will use components from Stream and Chatscope libraries.</p><p>Create a file <strong>Messages.tsx</strong> in the app directory with the following code:</p><pre>import { cn } from &quot;@/lib/utils&quot;;<br>import { useChannelStateContext } from &quot;stream-chat-react&quot;;<br>import { Message, MessageList } from &quot;@chatscope/chat-ui-kit-react&quot;;<br><br>export default function () {<br>  const { messages } = useChannelStateContext();<br>  return (<br>    &lt;MessageList&gt;<br>      {messages?.map((i, index: number) =&gt; (<br>        &lt;Message<br>          key={i.id}<br>          model={{<br>            position: &quot;normal&quot;,<br>            sender: i.user?.id,<br>            direction: &quot;incoming&quot;,<br>            message: `${i.user?.id}: ${i.text}`,<br>            sentTime: i.created_at?.toString(),<br>          }}<br>          className={cn(<br>            &quot;bg-white rounded text-black py-2 text-xs&quot;,<br>            index !== messages.length - 1 &amp;&amp; &quot;border-b&quot;<br>          )}<br>        /&gt;<br>      ))}<br>    &lt;/MessageList&gt;<br>  );<br>}</pre><p>Using the <strong>useChannelStateContext </strong>hook, you are able to retrieve the chat history, and listen to incoming messages. Using <strong>MessageList </strong>and<strong> Message </strong>Chatscope UI components you are able to mark all the messages as incoming to the channel, and map them with their sender’s information, message content, and timestamp.</p><p>Now, let’s build the route where users can view the chat history and post their message simultaneously. Update <strong>page.tsx</strong> file in the app directory with the following code:</p><pre>&quot;use client&quot;;<br><br>import { useState } from &quot;react&quot;;<br>import Messages from &quot;./Messages&quot;;<br>import { Button } from &quot;@/components/ui/button&quot;;<br>import { Channel, Chat } from &quot;stream-chat-react&quot;;<br>import { Textarea } from &quot;@/components/ui/textarea&quot;;<br><br>export default function () {<br>  const [channel, setChannel] = useState(null);<br>  const [chatClient, setChatClient] = useState(null);<br>  return (<br>    &lt;div className=&quot;flex max-w-[300px] flex-col gap-y-3 p-5&quot;&gt;<br>      &lt;div className=&quot;flex w-[300px] flex-col gap-y-3&quot;&gt;<br>        &lt;span className=&quot;border-b border-gray-100 font-semibold&quot;&gt;Chat&lt;/span&gt;<br>        {channel &amp;&amp; (<br>          &lt;Chat client={chatClient}&gt;<br>            &lt;Channel channel={channel}&gt;<br>              &lt;Messages /&gt;<br>            &lt;/Channel&gt;<br>          &lt;/Chat&gt;<br>        )}<br>        &lt;Textarea<br>          id=&quot;message_text&quot;<br>          name=&quot;message_text&quot;<br>          placeholder=&quot;Message...&quot;<br>          className=&quot;min-h-[100px] w-full&quot;<br>        /&gt;<br>        &lt;Button className=&quot;max-w-max&quot;&gt;<br>          Send Message &amp;rarr;<br>        &lt;/Button&gt;<br>      &lt;/div&gt;<br>    &lt;/div&gt;<br>  );<br>}</pre><p>The index route is now able to render the list of messages, and has a textarea element to accept input from a visitor. Before implementing the functionality to post a message, you will want to associate them with a unique id. Update the code with the following:</p><pre>&quot;use client&quot;;<br><br>import Messages from &quot;./Messages&quot;;<br>+ import { useEffect, useState } from &quot;react&quot;;<br>import { Button } from &quot;@/components/ui/button&quot;;<br>import { Channel, Chat } from &quot;stream-chat-react&quot;;<br>+ import { DevToken, StreamChat } from &quot;stream-chat&quot;;<br>import { Textarea } from &quot;@/components/ui/textarea&quot;;<br>+ import { generateUsername } from &quot;unique-username-generator&quot;;<br><br>export default function () {<br>  const [channel, setChannel] = useState(null);<br>  const [chatClient, setChatClient] = useState(null);<br>+ const loadChatClient = async () =&gt; {<br>+   const newChatClient = new StreamChat(<br>+     process.env.NEXT_PUBLIC_STREAM_API_KEY,<br>+     {<br>+       enableWSFallback: true,<br>+     }<br>+   );<br>+    if (newChatClient.user) await newChatClient.disconnectUser();<br>+   const localUser = localStorage.getItem(&quot;local_user&quot;);<br>+   if (!localUser) localStorage.setItem(&quot;local_user&quot;, generateUsername());<br>+   const id = localStorage.getItem(&quot;local_user&quot;);<br>+   const userToConnect = { id };<br>+   await newChatClient.connectUser(userToConnect, DevToken(userToConnect.id));<br>+   setChatClient(newChatClient);<br>+ };<br>+ useEffect(() =&gt; {<br>+   loadChatClient();<br>+ }, []);<br>  return (<br>    &lt;div className=&quot;flex max-w-[300px] flex-col gap-y-3 p-5&quot;&gt;<br>      {/* Rest of component as is */}<br>    &lt;/div&gt;<br>  );<br>}</pre><p>You have now made a one-time call to <strong>loadChatClient </strong>function, which enables Websocket connection to a messaging channel on Stream. It then disconnects any user (if) associated with the session. Finally, it generates a unique username for the visitor (if not found in localStorage<strong>)</strong>. These steps ensure that the visitor is identified uniquely.</p><p>To listen to messages in a particular channel, update the code with following:</p><pre>&quot;use client&quot;;<br><br>// Imports as is<br><br>export default function () {<br>  const [channel, setChannel] = useState(null);<br>  const [chatClient, setChatClient] = useState(null);<br>+  const watchChannel = () =&gt; {<br>+    const channel = chatClient.channel(&quot;messaging&quot;, &quot;livestreaming_chat&quot;, {<br>+      name: &quot;Live Streaming Chat&quot;,<br>+    });<br>+    channel.watch().then(() =&gt; setChannel(channel));<br>+  };<br>  const loadChatClient = async () =&gt; {<br>    const newChatClient = new StreamChat(<br>      process.env.NEXT_PUBLIC_STREAM_API_KEY,<br>      {<br>        enableWSFallback: true,<br>      }<br>    );<br>    if (newChatClient.user) await newChatClient.disconnectUser();<br>    const localUser = localStorage.getItem(&quot;local_user&quot;);<br>    if (!localUser) localStorage.setItem(&quot;local_user&quot;, generateUsername());<br>    const id = localStorage.getItem(&quot;local_user&quot;);<br>    const userToConnect = { id };<br>    await newChatClient.connectUser(userToConnect, DevToken(userToConnect.id));<br>    setChatClient(newChatClient);<br>  };<br>  useEffect(() =&gt; {<br>    loadChatClient();<br>  }, []);<br>+  useEffect(() =&gt; {<br>+    if (chatClient) watchChannel();<br>+  }, [chatClient]);<br>  return (<br>    &lt;div className=&quot;flex max-w-[300px] flex-col gap-y-3 p-5&quot;&gt;<br>      {/* Rest of component as is */}<br>    &lt;/div&gt;<br>  );<br>}</pre><p>The visitor is now connected to the <strong>Live Stream Chat</strong> channel, identified with <strong>livestreaming_chat</strong> unqiue id. Then, the <strong>watch() </strong>utility is used to listen and update the context with incoming messages.</p><p>To post a message in the channel, update the code with the following:</p><pre>&quot;use client&quot;;<br><br>// Imports as is<br><br>export default function () {<br>  // Variables, Hooks as is<br>  return (<br>    &lt;div className=&quot;flex max-w-[300px] flex-col gap-y-3 p-5&quot;&gt;<br>      &lt;div className=&quot;flex w-[300px] flex-col gap-y-3&quot;&gt;<br>        {/* Rest of component as is */}<br>        &lt;Button<br>          className=&quot;max-w-max&quot;<br>+          onClick={() =&gt; {<br>+            if (channel) {<br>+              channel.sendMessage({<br>+                text: document.getElementById(&quot;message_text&quot;).value,<br>+              });<br>+              document.getElementById(&quot;message_text&quot;).value = &quot;&quot;;<br>+            }<br>+          }}<br>        &gt;<br>          Send Message &amp;rarr;<br>        &lt;/Button&gt;<br>      &lt;/div&gt;<br>    &lt;/div&gt;<br>  );<br>}</pre><p>Using the <strong>sendMessage </strong>utility in the <strong>onClick</strong> event, visitors are able to post a message (in the channel). Neat!</p><p>You are now done with building the reactive chat interface that randomly generates a unique id for the visitor, renders a list of messages dynamically and allows them to post a message. Let’s deploy the Next.js application to Vercel.</p><h3>Deploy To Vercel</h3><p>The code is now ready to deploy to Vercel. Use the following steps to deploy:</p><ul><li>Start by creating a GitHub repository containing your app’s code.</li><li>Then, navigate to the Vercel Dashboard and create a <strong>New Project</strong>.</li><li>Link the new project to the GitHub repository you just created.</li><li>In <strong>Settings</strong>, update the <strong>Environment Variables</strong> to match those in your local .env file.</li><li>Click <strong>Deploy</strong>.</li></ul><h3>Ending Thoughts</h3><p>In conclusion, this tutorial provides a comprehensive guide on integrating Next.js and Stream to build a dynamic livestream chat experience. By granting users the ability to participate without authentication and ensuring granular control over their permissions, you have learnt how to make real-time interaction more accessible.</p><h3>More Information</h3><p>For more detailed insights, explore the references cited in this post.</p><ul><li><a href="https://github.com/rishi-raj-jain/my-chat-app">GitHub Repository</a></li><li><a href="https://github.com/chatscope/chat-ui-kit-react">Chatscope UI Kit</a></li><li><a href="https://getstream.io/chat/docs/react/authless_users/">Authless Users — Stream</a></li><li><a href="https://getstream.io/chat/docs/react/watch_channel/">Watching a Channel — Stream</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e23ca1fe29ec" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Custom Service Worker in any app with esbuild]]></title>
            <link>https://medium.com/@rishi-raj-jain/custom-service-worker-in-any-app-with-esbuild-6b5a8a20fa12?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/6b5a8a20fa12</guid>
            <category><![CDATA[service-worker]]></category>
            <category><![CDATA[web]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[performance]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Sun, 23 Jan 2022 11:52:47 GMT</pubDate>
            <atom:updated>2022-01-26T05:37:12.742Z</atom:updated>
            <content:encoded><![CDATA[<p>Lately I’ve been exploring use of Service Workers in web apps. Use cases include <a href="https://docs.layer0.co/guides/prefetching">prefetching content at scale</a>, <a href="https://www.builder.io/blog/how-we-cut-99-percent-js-with-qwik-and-partytown">speeding up your website</a>, <a href="https://dexecure.com/blog/beyond-offline-other-interesting-use-cases-for-service-workers-2/#automated-failover">handling failovers</a> and possibly many more. One of the problems that I continuously face is the need to adjust my service worker according to the framework, well no more. Let’s see what gold I dig up last week below 👇🏻</p><h3>Where it started?</h3><p>I was trying to set up a custom service worker with Angular 11, and the guide <a href="https://angular.io/guide/service-worker-getting-started">https://angular.io/guide/service-worker-getting-started</a> asked me to install angular/pwa module, read <a href="https://angular.io/guide/service-worker-communications">https://angular.io/guide/service-worker-communications</a> and then adapt to the angular’s way of creating service worker. Well, seemed like a lot of work to me, so I set out to create a framework-agnostic process for adding a custom service worker. Reflecting back, I think it was <a href="https://rishi.app/blog/enabling-service-worker-with-vue-2-and-vue-3">fairly easier with Vue 2 and 3</a>, yet felt like a hacky way.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iY0CHjoKab0xO52Kt3nPiw.jpeg" /><figcaption>(What is a <a href="https://blog.sessionstack.com/how-javascript-works-service-workers-their-life-cycle-and-use-cases-52b19ad98b58">Service Worker</a>?)</figcaption></figure><h3>Show me the solution!</h3><p>Step 0. Install <a href="https://esbuild.github.io/getting-started/#install-esbuild"><strong>esbuild</strong></a></p><p>Step 1. Create <strong>workbox-config.js:</strong></p><p>Learn more about what’s there in workbox config here:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/bc23b80c6900e5322e11d3a6cffc99ad/href">https://medium.com/media/bc23b80c6900e5322e11d3a6cffc99ad/href</a></iframe><p>Step 2. Create <strong>compile_sw.js:</strong></p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d149178807ef7d2b23aad0e8dabbe480/href">https://medium.com/media/d149178807ef7d2b23aad0e8dabbe480/href</a></iframe><p>Step 3.<strong> Compile and inject:</strong></p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a9cce7f281b0c348c8bb319256d1c2d3/href">https://medium.com/media/a9cce7f281b0c348c8bb319256d1c2d3/href</a></iframe><p>Step 4. <strong>Register the Service Worker</strong> ✨</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/0299f8e534133b0191950811a18f03e2/href">https://medium.com/media/0299f8e534133b0191950811a18f03e2/href</a></iframe><p>That’s all, enjoy! I hope this helps!</p><h3><strong>References</strong></h3><ul><li><a href="https://developers.google.com/web/tools/workbox/modules/workbox-webpack-plugin">https://developers.google.com/web/tools/workbox/modules/workbox-webpack-plugin</a></li><li><a href="https://github.com/evanw/esbuild/issues/466">https://github.com/evanw/esbuild/issues/466</a></li><li><a href="https://github.com/evanw/esbuild/issues/12">https://github.com/evanw/esbuild/issues/12</a></li><li><a href="https://github.com/evanw/esbuild/issues/69">https://github.com/evanw/esbuild/issues/69</a></li><li><a href="https://github.com/evanw/esbuild/issues/312">https://github.com/evanw/esbuild/issues/312</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6b5a8a20fa12" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From delivering content to eXperience — How Layer0 has changed my eXperience as a Front-End…]]></title>
            <link>https://medium.com/@rishi-raj-jain/from-delivering-content-to-experience-how-layer0-has-changed-my-experience-as-a-front-end-191d48437e72?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/191d48437e72</guid>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[angular]]></category>
            <category><![CDATA[front-end-development]]></category>
            <category><![CDATA[ui-design]]></category>
            <category><![CDATA[server-side-rendering]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Sun, 18 Apr 2021 10:10:43 GMT</pubDate>
            <atom:updated>2021-04-22T16:12:28.866Z</atom:updated>
            <content:encoded><![CDATA[<h3>From delivering content to <strong>eXperience — </strong>How Layer0 has changed my eXperience as a Front-End D̶e̶v̶e̶l̶o̶p̶e̶r Engineer</h3><p>I’ve been doing Front-End Development for more than a year now and recently have been exploring the domain of <strong>Front</strong>-<strong>End Engineering </strong>where the creators go beyond development and focus on usability &amp; user experience. I’ve been a fan of Vercel ever since I started developing with Next.js and now Layer0 has certainly made it’s way to the top of the list. This blog will keep you apprised of my learnings from Next.js Conf, my experience with Vercel, what convinced me of Layer0 and my experience with deploying an application charged up by the Fastest Backend for Frontend(BFF), Layer0.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ch9Alci61wF75Kb9oFracg.png" /></figure><h3>My learnings from Next.js Conf</h3><p>I used to think about what possibly could one learn from a conference? But after attending the <a href="https://nextjs.org/2020/conf">Next.js very first global user conference</a>, I’m now more than excited to attend and keep learning from such conferences. The developers have brought several jaw-dropping features in Next.js 10 such as:</p><ul><li><a href="https://nextjs.org/blog/next-10#built-in-image-component-and-automatic-image-optimization"><strong>Built-in Image Component and Automatic Image Optimization</strong></a><strong>: </strong>Allowing one to sit back and let next.js take care of improving the website performance by resizing, optimising, and serving images in modern formats &amp; lazy loading them in run time.</li><li><a href="https://nextjs.org/blog/next-10#automatic-resolving-of-href"><strong>Automatic Resolving of </strong></a><strong>`href`: </strong>Where developers won’t be needing the as` prop to configure how would their routes display in the URL bar.</li><li><a href="https://nextjs.org/blog/next-10#getstaticprops--getserversideprops-fast-refresh"><strong>getStaticProps / getServerSideProps Fast Refresh:</strong></a> While editing data fetching methods, the function is re-run, and now reloading the page won’t be a requirement 🎉.</li></ul><p>But for me, here’s ‘the’ key takeaway from the conference →</p><blockquote><strong>Google’s recent announcement on </strong><a href="https://webmasters.googleblog.com/2020/05/evaluating-page-experience.html"><strong>Evaluating page experience for a better web</strong></a><strong>, according to which apart from the content, performance of websites would now be a major factor in prioritising them on google search results!</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/938/1*2dLd5w6Q9MrmfAQfLoJ4Zw.png" /><figcaption>(<a href="https://webmasters.googleblog.com/2020/05/evaluating-page-experience.html">Page experience ranking factors</a>)</figcaption></figure><blockquote>This will now push each and every developer around the globe to deliver sites with a greater user experience to stay at the top &amp; maintain their rankings.</blockquote><h3><strong>My Experience with Vercel (formerly Zeit) [The platform made for Next.js]</strong></h3><p>From static websites to hosting dynamic websites built with Angular, Vue &amp; React, I’ve been using Vercel for almost everything. Maybe why I keep using it is because users get free unlimited domains to deploy on, with zero configuration required. With easy deployments from Vercel CLI, the process of deploying couldn’t get any smoother.</p><p>But with various kinds of deployments, I came to know of some issues with Vercel hosting. For the same website that I could open on my internet, I was frequently told that <a href="https://www.facebook.com/groups/DevCDelhiNCR/permalink/1211002872566803/?comment_id=1211049259228831&amp;reply_comment_id=1212322355768188">“The website displays an error message and doesn’t load”</a> or <a href="https://www.facebook.com/groups/DevCDelhiNCR/permalink/1211002872566803/?comment_id=1211049259228831&amp;reply_comment_id=1212322355768188">“The website couldn&#39;t load”</a>. Once in a while, such messages can be ignored but here was the deal-breaker which led me to Layer0, that I had to host an <a href="https://angular.io/guide/universal">Angular SSR</a> <a href="https://github.com/rishi-raj-jain/Blogging">application </a>which <a href="https://github.com/vercel/ng-deploy-vercel/issues/9">Vercel currently doesn’t support</a>! 😲 (Shocking, right!?)</p><h3><strong>What did Layer0 have to offer?</strong></h3><p>I came to know of Layer0 at the <a href="https://nextjs.org/2020/conf">Next.js Conf</a> and then discovered the platform that promises instant website load, a very smooth deploying experience, all in one platform to develop, deploy, preview, run, monitor, experiment on, and defend one’s favorite frontend framework!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pAfZAtErrRRaAI3uTQscNQ.png" /><figcaption>(<a href="https://www.moovweb.com/products">CDN-as-JavaScript puts data into the browser 5 seconds <em>before</em> it’s needed)</a></figcaption></figure><blockquote>“Layer0 <strong>extends</strong> the capabilities of a traditional CDN by not only hosting your static content, but also providing <strong>server-side rendering</strong> for progressive web applications as well as <strong>caching both your APIs and HTML at the network edge</strong> to provide your users with the <strong>fastest browsing experience</strong>.” ~ <a href="https://docs.layer0.co/guides/overview#section_what_is_an_layer0_">Layer0 Docs</a></blockquote><h3>How is it any different from a Content Delivery Network (CDN)?</h3><p>With eCommerce sites becoming more interactive &amp; dynamic, creating layouts from JSON that are <strong>not</strong> cached by CDNs is a factor that affects the performance of several websites being built around the world.</p><p>eCommerce websites with such CDNs result in larger <a href="https://web.dev/fcp/">First Contentful Paint(FCP)</a> one of the core web vitals listed in Google’s recent announcement on-page experience, henceforth deteriorating the user experience. This leads to customers waiting for product data that needs to be pulled from the database even before the browser can render the page on the screen!</p><p>This meant that I could definitely host my sample Angular SSR blogging application in no time, and was very keen to test their claim of delivering sub second page load-times. 🤓</p><h3>Did I even have to do a thing?</h3><p>So the scenario is that I was already done developing my Angular SSR application (<a href="https://github.com/rishi-raj-jain/Blogging">Github Link</a>) and had now to deploy it via <a href="https://app.layer0.co/">app.layer.co</a>. I just went through their documentation for deploying Angular @ <a href="https://docs.layer0.co/guides/angular">https://docs.layer0.co/guides/angular</a> and ran these 2 commands to integrate layer0 with my existing angular project:</p><pre>npm install -g <a href="http://twitter.com/layer0/cli">@layer0/cli</a><br>layer0 init</pre><p>and was done with the integration part! Woah!</p><p>After doing a sample run via layer0 run , I tested the app locally exactly as it will be uploaded to the Layer0 cloud using serverless-offline via layer0 build &amp;&amp; layer0 run --production and to my happiness, it worked flawlessly!</p><p>To get it running at Layer0’s edge network, I ran layer0 deploy , and Voila! I’m now the owner of most probably the fastest site in the world xD</p><p>Don’t believe me? Have a look at for yourself @ <a href="https://rishi-raj-jain-my-app-default.moovweb-edge.io/">https://rishi-raj-jain-my-app-default.moovweb-edge.io/</a></p><h3>The satisfying eXperience</h3><p>I ran Google Lighthouse and believe me I was awestruck by the website’s performance. I switched from my Wi-Fi to streaming on my mobile’s data network but it continued to surprise me with just 162ms of First Contentful Paint (FCP) [basically when a user is able to process input from the webpage], estimated 0 total blocking time (TBT), and with 211.1 ms as the time when the entire DOMContent gets loaded!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9mnFbIQfDT_LhDeeNAvCRQ.png" /><figcaption>(Performance Metrics of <a href="https://rishi-raj-jain-my-app-default.moovweb-edge.io/">https://rishi-raj-jain-my-app-default.moovweb-edge.io/</a>)</figcaption></figure><h3>Conclusion</h3><p>A faster website would attract more users to it. Simple, right? While this seems to be fair what if I say a second of delay would be fine. Would it be? if yes, here’s something that needs your attention:</p><p>“Amazon’s calculated that a page load slowdown of just one second could cost it $1.6 billion in sales each year. Google has calculated that by slowing its search results by just four tenths of a second they could lose 8 million searches per day–meaning they’d serve up many millions fewer online adverts.” - <a href="https://www.fastcompany.com/1825005/how-one-second-could-cost-amazon-16-billion-sales">(KIT EATON, Fast Company)</a></p><p>It might be to one’s surprise, but certainly even every sub second of delay matters! Seems like Layer0 is doing a fine job in providing a better developer experience by their all-in-one platform for deployment, and a greater user experience pertaining to the fastest loads I’ve ever witnessed.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=191d48437e72" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Dynamic Markdown Blogs in Next.js/React]]></title>
            <link>https://medium.com/@rishi-raj-jain/dynamic-markdown-blogs-in-next-js-react-c0ed10d2d8a2?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/c0ed10d2d8a2</guid>
            <category><![CDATA[blog]]></category>
            <category><![CDATA[serverless]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[nextjs]]></category>
            <category><![CDATA[react]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Sat, 09 Jan 2021 20:29:30 GMT</pubDate>
            <atom:updated>2021-10-27T10:01:32.522Z</atom:updated>
            <content:encoded><![CDATA[<h3>Dynamic Markdown Blogs in Next.js/React using gray-matter, react-markdown and react-syntax-highlighter</h3><p>You are looking for a sustainable, scalable solution to create dynamic blogs whether it be for your website, or someone else’s, then react-markdown is the solution for it. This blog assumes, obviously, you are working with Next.js/React.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DZqkBELtxVeDk-HJNXUfNg.png" /></figure><p>To get started with, create a directory say ‘_snippets’ and in this directory create a sample markdown file, say <strong>‘Dynamic Markdown Blogs in Next.js using gray-matter, react-markdown and react-syntax-highlighter.md’</strong>. The slug of the route for the particular blog would be derived from the markdown file name itself.</p><p>Here’s the sample markdown for you to copy:</p><pre>---<br>title: &#39;Dynamic Markdown Blogs in Next.js/React using gray-matter, react-markdown and react-syntax-highlighter&#39;<br>excerpt: &#39;Implementing markdown blogs in Next.js/React with remark-markdown in just 5 minutes. Thinking of using strings? Nah, they are too old, use markdown. Export statically to achieve great perfomance.&#39;<br>date: &#39;2021-01-06&#39;<br>---<br><br>You are looking for a sustainable, scalable solution to create dynamic blogs whether it be for your website, or someone else&#39;s, then react-markdown is the solution for it. This blog assumes, obviously, you are working with Next.js/React.</pre><p>Done with the content 🚀 &amp; proceed by executing the following on your terminal:</p><pre>npm i react-syntax-highlighter gray-matter react-markdown</pre><p>Done with the installation 🚀, proceed by creating the following component:</p><pre>// File: /components/MarkdownHighlight.js<br><br>import React from &#39;react&#39;<br>import { Prism as SyntaxHighlighter } from &#39;react-syntax-highlighter&#39;<br>import {<br>  materialDark,<br>  materialLight,<br>} from &#39;react-syntax-highlighter/dist/cjs/styles/prism&#39;<br><br>const MarkdownHighlight = ({ darkMode, language, value }) =&gt; {<br>  return (<br>    &lt;SyntaxHighlighter<br>      language={language}<br>      style={darkMode ? materialLight : materialDark}<br>    &gt;<br>      {value}<br>    &lt;/SyntaxHighlighter&gt;<br>  )<br>}<br><br>export default MarkdownHighlight</pre><p>This component is gonna take care of highlighting your markdown code blocks whether html, css, js or markdown. Browse <strong>react-syntax-highlighter/dist/cjs/styles/prism</strong> for exploring different kind of filters such as {coy}, {dark} apart from {materialDark}, {materialLight}.</p><p>The code below will be handling formation of blog data from the markdown files. Explanations for each function has been added.</p><pre>// File: /lib/markdown.js<br><br>import { join } from &#39;path&#39;<br>import fs from &#39;fs&#39;<br>import matter from &#39;gray-matter&#39;<br><br>// Directory of snippets<br>const pagesDirectory = join(process.cwd(), &#39;_snippets&#39;)<br><br>// Form slugs from the markdown names<br>export function getSlugsFromDirectory(dir) {<br>  return fs.readdirSync(dir)<br>}<br><br>/**<br> * Gets the contents of a file<br> * The gray-matter (metadata at the top of the file) will be<br> * added to the item object, the content will be in<br> * item.content and the file name (slug) will be in item.slug.<br> */<br>export function getBySlug(dir, slug, fields = []) {<br>  const realSlug = slug.replace(/\.md$/, &#39;&#39;)<br>  const fullPath = join(dir, `${realSlug}.md`)<br>  const fileContents = fs.readFileSync(fullPath, &#39;utf8&#39;)<br>  const { data, content } = matter(fileContents)<br>  const items = {}<br>  fields.forEach((field) =&gt; {<br>    if (field === &#39;slug&#39;) {<br>      items[field] = realSlug<br>    }<br>    if (field === &#39;content&#39;) {<br>      items[field] = content<br>    }<br>    if (data[field]) {<br>      items[field] = data[field]<br>    }<br>  })<br>  return items<br>}<br><br>// Returns contents of a page in the _snippets directory<br>export function getPageContentBySlug(slug, fields = []) {<br>  return getBySlug(pagesDirectory, slug, fields)<br>}<br><br>// Returns pages in the _snippets directory<br>export function getAllSnippets(fields = []) {<br>  const slugs = getSlugsFromDirectory(pagesDirectory)<br>  const pages = slugs.map((slug) =&gt; getPageContentBySlug(slug, fields))<br>  return pages<br>}</pre><p>Done with the creating helping functions 🚀, let’s proceed by creating the dynamic blog page:</p><pre>// File: /pages/snippet/[slug].js<br><br>import { useRouter } from &#39;next/router&#39;<br>import ReactMarkdown from &#39;react-markdown&#39;<br>import MarkdownHighlight from &#39;@/components/MarkdownHighlight&#39;<br>import { getAllSnippets, getPageContentBySlug } from &#39;@/lib/markdown&#39;<br><br>const Snippet = ({ page, darkMode }) =&gt; {<br>  const router = new useRouter()<br>  return router.isFallback ? (<br>    &lt;div&gt;Loading...&lt;/div&gt;<br>  ) : (<br>    &lt;div className=&quot;container&quot;&gt;<br>      &lt;h1&gt;{page.title}&lt;/h1&gt;<br>      &lt;div className={styles[&#39;mt-3&#39;]}&gt;<br>        &lt;div className={darkMode ? &#39;text-white&#39; : &#39;text-dark&#39;}&gt;<br>          &lt;ReactMarkdown<br>            source={page.content}<br>            renderers={{<br>              code: ({ language, value }) =&gt; {<br>                /* Automatically takes in the language &amp; value from the markdown file when: ```&lt;html/css/js&gt;<br>                  Content here<br>                  ``` in the markdown file*/<br>                return (<br>                  &lt;MarkdownHighlight<br>                    language={language}<br>                    value={value}<br>                    darkMode={darkMode}<br>                  /&gt;<br>                )<br>              },<br>            }}<br>          /&gt;<br>        &lt;/div&gt;<br>      &lt;/div&gt;<br>    &lt;/div&gt;<br>  )<br>}<br><br>export default Snippet<br><br>export async function getStaticProps({ params }) {<br>  const { slug } = params<br>  const page = getPageContentBySlug(slug, [<br>    &#39;title&#39;,<br>    &#39;image&#39;,<br>    &#39;slug&#39;,<br>    &#39;content&#39;,<br>    &#39;date&#39;,<br>  ])<br>  return {<br>    props: {<br>      page: {<br>        ...page,<br>        markdown: page.content,<br>      },<br>    },<br>  }<br>}<br><br>export async function getStaticPaths() {<br>  const posts = getAllSnippets([&#39;slug&#39;])<br>  const paths = posts.map(({ slug }) =&gt; ({<br>    params: {<br>      slug,<br>    },<br>  }))<br>  return {<br>    paths,<br>    fallback: &#39;blocking&#39;<br>  }<br>}</pre><p><strong>All set! Hope you liked it :)</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c0ed10d2d8a2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Case study: Redesigning Snapchat money for gen-Z]]></title>
            <link>https://medium.com/design-bootcamp/redesigning-snapcash-for-gen-z-introducing-snapchat-money-for-snapchat-india-fea7644ca630?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/fea7644ca630</guid>
            <category><![CDATA[design-thinking]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[snapchat]]></category>
            <category><![CDATA[case-study]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Fri, 20 Nov 2020 09:52:47 GMT</pubDate>
            <atom:updated>2020-12-07T14:50:17.878Z</atom:updated>
            <content:encoded><![CDATA[<h4>Introducing Snapchat Money for Snapchat India</h4><h3>Too long; didn’t read.</h3><p>This is a case study of designing an <a href="https://www.productplan.com/glossary/minimum-viable-feature/">MVF (Minimal Viable Feature)</a> from scratch in 2 days — A <a href="https://www.ndtv.com/india-news/air-quality-in-delhi-adjoining-areas-falls-to-severe-category-thick-smog-reduces-visibility-2325268">pollution</a> packed journey starting from doing an assignment to leaving that assignment and going through intense research and analysis to re-designing <a href="https://www.snap.com/en-US/news/post/introducing-snapcash">Snapcash</a> for Generation-Z (Age: 13–18) to enable money transfers via SnapChat Money, based on rational hypothesis &amp; modern <a href="https://www.interaction-design.org/literature/topics/design-thinking?ep=ug0">design thinking</a>.</p><h4><strong>Disclaimer</strong></h4><p>The proposed feature is in no way an official feature rolled out by Snapchat; the work is being done solely as a student learning design.</p><p>While designing for the task, I did not intend to mar Snapchat’s online image. Apart from that, this is not being done for any kind of business, technological, or another resource-based benefit. I’ve tried to follow each and every <a href="https://storage.googleapis.com/snap-inc/brand-guidelines/snapchat-brand-standards.pdf">guideline</a> issued by Snapchat for whenever their design elements are being used. No copyright infringement is intended.</p><h3>The Task</h3><blockquote>Imagine, you were recently hired in the <a href="https://www.snapchat.com/">Snapchat</a> Design team. The team plans to release a new feature “SnapChat Money” in the current Snapchat app for India. They will introduce a wallet (not UPI) into which money can be added from external sources (Debit card/Credit Card/UPI/Net Banking) and then can be used for intra-Snapchat transfers,i.e., Snapchat Money can be transferred between Snapchat users and you have been appointed to lead and execute the entire design.</blockquote><h3>The discovery of Snapcash</h3><p>After reading the task, it struck me that this is something Snapchat would’ve also considered, and so to satisfy my curiosity, I googled. I was shocked to see that Snapchat in 2015 had already introduced such a feature, <a href="https://www.snap.com/en-US/news/post/introducing-snapcash">Snapcash</a>, which did exactly as intended but was scrapped due to its misuse! People were to enter their debit card linked with <a href="https://squareup.com/us/en">Square</a>, then just type say $5 and send the message to the recipient &amp; Voila! Here’s a video that explains Snapcash quite well →</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FR2KRwgcxUHk%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DR2KRwgcxUHk&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FR2KRwgcxUHk%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/229fda7dd34c3c2e12f4775751493985/href">https://medium.com/media/229fda7dd34c3c2e12f4775751493985/href</a></iframe><p>Even though I see it as a strategic move by Snapchat to enter the payment market and take advantage of its humongous user base, the feature “ended up as a way to pay adult performers for private content over Snapchat, not just a way to split bills with friends.” ~ (Josh, <a href="https://techcrunch.com/2018/07/22/snapcashing-out/">TechCrunch</a>)</p><p>I think even Snapchat didn’t expect their medium to be used for such purposes; hence they had to put an <a href="https://techcrunch.com/2018/07/22/snapcashing-out/">end to their feature</a>. While this doesn’t give out much about Gen-Z, this was insightful as this helped me base my assumptions before I ended up designing something that’d allow such use of the medium. It also helps me prove that this is a potential area to work for and gives the idea of how such a feature was to be designed while maintaining the uniqueness of Snapchat in it.</p><h3>Problem Statement</h3><p>Sometimes motivation is all you need. Upon searching the `famous product designer quotes,` this stood out, “The only important thing about design is how it relates to people.” ~ <a href="https://en.wikipedia.org/wiki/Victor_Papanek">Victor Papanek</a>. Well, now was the time to talk to a Gen-Z who use(s/d) Snapchat, understand how do they conduct their payment activities, the problems they face in the process, listen to some mishappenings (if you will), and what features/services/elements they liked about specific payment applications. At last, some insight into their Snapchat usage is deemed sensible.</p><h3>User Analysis</h3><p>Being a junior at a university, sometimes you are nothing but serious (Wait for it), I asked one of my friends if they would Snapchat for money, and the instant reply that I got “If I were this serious, I’d not rather use Snapchat in the first place.” But considering that my friend aka a 20-year-old boy, didn’t belong to the targeted Gen-Z (13–18 years) users :P, I decided to settle with the observation that the payment activity needed to have a fun component to be perceived as a Snapchat feature and not an entirely different product.</p><p>To understand the usefulness of such a feature, as already said, I needed to understand the usage of Snapchat among 13–18 years old and their payment activities. So, I did two telephonic interviews with my younger cousins <a href="https://drive.google.com/drive/folders/1VkTzaRQTLHs1JCS6QFMsOMeiCyo9cBhR?usp=sharing">(Listen to the recordings here)</a> I focused on the following set of questions:</p><ul><li>Have you ever performed an electronic payment? (If the answer to the above question was yes, it was followed up by questions below, otherwise I intended to learn about the reasons that kept them from doing the same)</li><li>If yes, which medium did you use, and why?</li><li>On average, what’s the amount you pay/request per transaction?</li><li>Has there been an incident where something unexpected happened while performing a payment? Say you ever sent someone money by accident instead of requesting money from them or vice versa!?</li><li>How often do you use Snapchat?</li><li>What’s the primary reason you use Snapchat for?</li></ul><h3>Insights</h3><p>As someone who has not used a smartphone till class 11th, it was nice to see that Gen-Z is becoming aware &amp; taking up the responsibility at an early stage of life. After having those two calls, here are my learnings of their experiences of conducting payment activities:</p><ul><li><strong>Pain Point #1</strong>: U<strong>nexpected behavior of the transaction</strong> — the user is faced with confusion and couldn’t make sense of what happened when a transaction gets <strong>paid successfully &amp; refunded</strong> and is <strong>not received</strong> at the other end as well! And to now make sense of what has happened, there is no initiative from the system to detect such a fault, rather the support of the company X has to be contacted.</li><li><strong>Observation #1</strong>: — Though this is a very minute observation, the user assumed PhonePe &amp; Paytm to be the reliable sources of conducting payment activities as they were famous &amp; were already present in their system. This might also define the usability of Snapchat Money as it’d add to the users’ subconscious mind that as they rely on Snapchat to share their moments with others, paying via Snapchat would be reliable too.</li></ul><h3>Edge Cases</h3><p>The picture is clear now. But what about edge cases? Though easy to think, documentation is essential, and thus, here goes:</p><ul><li>Users trying to transfer money but with insufficient funds — Requires addition of money to the wallet via external sources(Debit card/Credit Card/UPI/Net Banking)</li><li>Users lose access to the internet while processing transaction — As the default behavior of transaction is, an incomplete transaction would result in no deduction of Snapchat Money between Sender(s) &amp; Receiver(s)</li><li>A user transferred the money while the system fails to acknowledge it at the other end.</li><li>While adding money to the wallet from the bank, while it shows a deduction of the amount from the bank, the money doesn’t show up in the wallet.</li></ul><h3>Security Concerns</h3><p>While this seems something to be on the implementation part, it is essential to keep the user feel safe by using the native elements of the application itself. Apart from verifying the Gen-Z users via Aadhar Card, it becomes necessary to have a Security PIN be entered for each and every Intra-Snapchat Transfer to keep users feel satisfied &amp; safe that their security is the primary concern.</p><h3>The Solution</h3><p>The following solution is based on the following assumptions:</p><ul><li>My cousins didn’t lie xD</li><li>Users are comfortable using Snapchat</li><li><a href="https://techcrunch.com/2018/07/22/snapcashing-out/">Users don’t intend to misuse such a service</a></li><li>Users aren’t in haste to skip the instructions before using the service</li><li>Users do own at least an external source of money (Debit card/Credit Card/UPI/Net Banking)</li><li>The user is an Indian</li><li>The user has viable access to the internet</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LGHLpMFCuz3Xv9VOztK5cg.png" /><figcaption>(Information Architecture 1st Iteration | Snapchat Money)</figcaption></figure><p>The following was my first iteration over Information Architecture, where I intuitively created the organization, but this seemed to be an overkill for `Simplifying money transfers` so after putting some serious thought into it, I came up with the following:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0eVrTqQeCqrjVUHXuC1LlQ.png" /><figcaption>(Information Architecture 2nd Iteration | Snapchat Money)</figcaption></figure><p>The second iteration is better as it actually simplifies the flow. The user(s) either pays/requests to/from a specific user(s) by chat, or switch to the account section, access their wallet, and proceed with their money pay/request or view their transaction history.</p><p>Considering all the edge cases, this is the user flow (<a href="https://miro.com/app/board/o9J_leimxAI=/">view on Miro</a>) I came up with →</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*x_4vd4qH_ZwtwPPuzwLnhg.jpeg" /></figure><h3>Wireframe Sketches</h3><p><strong>Low Fidelity</strong></p><p>Though I’m still developing my drawing skills, here the low fidelity represents a simple scenario where the user intends to make a payment after adding money to the wallet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jmBf829VraY50PEphMQWmg.jpeg" /><figcaption>(Step 1: Adding money to Snapchat Wallet)</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*otReHKj0j_Gp8WnbAQ4A1w.jpeg" /><figcaption>(Step 2: Paying money to friends via Snapchat Wallet)</figcaption></figure><h3>Tada… Introducing Snapchat Money!</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/820/1*q8gI8O3G2t9oKo1UxvElWw.png" /><figcaption>(Snapchat Money, Promotional Banner) [Background courtesy <a href="https://support.snapchat.com/en-US">Snapchat Support</a>]</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MRCA6v01m9phIql1e0lrhw.png" /><figcaption>(Money Transfers Via Snapchat Money)</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/912/1*lJnN2cUpysYoGIyjyVohNA.png" /><figcaption>(Account, Wallet Screens — Snapchat Money)</figcaption></figure><h3>Conclusion</h3><p>Another fun project! I felt this was a quite realistic problem and needs to be researched. Also while solving I worked with that one task, I chose in my mind, so the design may be a bit biased towards this one workflow, which may cause issues during the actual design and development of the product. Also, the article is a bit heavy on text, so if you have made it this far… I appreciate your patience. <strong>Thanks for reading :)</strong></p><h3>References</h3><ul><li><a href="https://medium.com/usabilitygeek/240-hours-of-ux-nutanix-design-competition-f3a98123ccdb">https://medium.com/usabilitygeek/240-hours-of-ux-nutanix-design-competition-f3a98123ccdb</a></li><li><a href="https://uxdesign.cc/ux-case-study-venmos-pay-request-money-feature-b3499952e41e">https://uxdesign.cc/ux-case-study-venmos-pay-request-money-feature-b3499952e41e</a></li><li><a href="https://www.snap.com/en-US/news/post/introducing-snapcash">https://www.snap.com/en-US/news/post/introducing-snapcash</a></li><li><a href="https://techcrunch.com/2018/07/22/snapcashing-out/">https://techcrunch.com/2018/07/22/snapcashing-out/</a></li><li><a href="https://uxplanet.org/information-architecture-basics-for-designers-b5d43df62e20">https://uxplanet.org/information-architecture-basics-for-designers-b5d43df62e20</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fea7644ca630" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/redesigning-snapcash-for-gen-z-introducing-snapchat-money-for-snapchat-india-fea7644ca630">Case study: Redesigning Snapchat money for gen-Z</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Case study: A UX-centered analysis of information sharing and privacy over WhatsApp]]></title>
            <link>https://medium.com/design-bootcamp/a-ux-centered-analysis-of-information-sharing-and-privacy-over-whatsapp-d1b65da18e7?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/d1b65da18e7</guid>
            <category><![CDATA[ux-research]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[case-study]]></category>
            <category><![CDATA[usability]]></category>
            <category><![CDATA[usability-heuristics]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Tue, 20 Oct 2020 17:17:08 GMT</pubDate>
            <atom:updated>2021-02-20T02:48:30.603Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SVc8fBl74ZJZbYPshVmjWg.png" /></figure><p>In my mid-semester examination, I had this problem statement which needed me to critically review two purposes for which WhatsApp is used but is not the best solution/ wasn’t intended for. (Well, I went beyond two :P)</p><p>WhatsApp has more than <strong>2 billion active users</strong> in 180 countries. WhatsApp started as an alternative to SMS and now they support sending and receiving a variety of media: text, photos, videos, documents, and location, as well as voice calls. Certainly, they’ve delivered a great User Experience to make their application stand out!</p><blockquote><em>Disclaimer</em></blockquote><blockquote><em>All of the observations, critiques, and suggestions are solely based on my comprehension and experience of using WhatsApp. While writing the critique, I did not consider any kind of business, technological, or other resource-based constraints.</em></blockquote><blockquote><em>The evaluation is part of my learning as a student of Design and is in no way an attempt to mar WhatsApp’s online image. No copyright infringement intended.</em></blockquote><h3>Information Sharing on WhatsApp is done via</h3><ol><li>Text Messages: Communicating over text,</li><li>Images: Capture/Attach images to communicate,</li><li>Audios: Record/Attach audios to communicate,</li><li>Videos: Record/Attach videos to communicate,</li><li>Documents: Share various kinds of document,</li><li>Forwarding: Forward messages from an individual or group chat to another individual or group chat,</li><li>Status: Share text, photos, videos, and animated GIFs that disappear after 24 hours,</li><li>Contacts: Share contact(s) with a click.</li></ol><h3>Severity Index</h3><p>I’ve used the severity index to rate and prioritize the problems based on my observations. This helped us identify and segregate the ‘major’ and ‘minor’ issues on a more comprehensive quantifiable scale of 0–4. Higher the severity index, severe the issue.</p><h3>Nielsen’s Heuristics</h3><p>To review, I used Nielsen’s Heuristics. These heuristics are general principles for interaction design. <a href="https://www.nngroup.com/topic/heuristic-evaluation/">Heuristic evaluation</a> (Nielsen and Molich, 1990; Nielsen 1994) is a <a href="http://www.nngroup.com/books/usability-engineering/">usability engineering</a> method for finding the usability problems in a user interface design so that they can be attended to as part of an iterative design process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/888/0*6chA7e6y5rCU0X1A.jpg" /><figcaption>(Image by <a href="https://www.interaction-design.org/literature/article/how-to-conduct-a-heuristic-evaluation-for-usability-in-hci-and-information-visualization">Interaction Design Foundation</a>)</figcaption></figure><h3>Reviewing the Experience</h3><p>Based on the heuristics, I found 15 usability issues in WhatsApp. I pointed out the major violated heuristic and graded their severity according to the severity index.</p><h4>Issue#1</h4><p>Unable to edit sent messages</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zq9ct_7wmJ50xsSpPSAE3A.jpeg" /></figure><p><strong>Severity: </strong>4</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>A user can not edit a sent message. To send the edited text, one has to delete the message and then send it again after making the desired changes. The usage frequency of this feature would definitely high and with more and more people switching to WhatsApp, this would help decrease the load of messages accompanying correction of the other.</p><h4>Issue#2</h4><p>Can not pin more than 3 chats</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/264/1*1AssHxUlO0nxP1m_Mm8fFA.png" /><figcaption>(Image by <a href="https://gadgets.ndtv.com/apps/news/whatsapp-iphone-update-pinned-chats-share-all-files-delete-multiple-photos-1725672">Gadgets360</a>)</figcaption></figure><p><strong>Severity: </strong>4</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>A user can not pin more than 3 chats. This limits the amount<br>of chats on the top, and now one has to go through the pain of <br>either searching for the chat on the Search bar or scrolling through <br>numerous chats until stumbles across the particular window. With such a huge user base, professionals would certainly be looking out for this.</p><h4>Issue#3</h4><p>Can not pin a message on a group</p><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>Any member (including admin) can not pin a message on the group. The purpose that the message is read at least once by every member, especially for those joining in thereafter is defeated. With such a huge user base, professionals would certainly be looking out for this too!</p><h4>Issue#4</h4><p>Groups/Broadcasts can not contain more than 256 members</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*b31hiO4ynbDLRrXWEFF4aQ.png" /><figcaption>(Image by <a href="https://www.bgr.in/news/whatsapp-now-allows-over-256-users-in-a-group-heres-a-quick-way-to-cross-the-user-limit-525042/">BGR.in</a>)</figcaption></figure><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>A group or a broadcast can not have more than 256 members. This gets really frustrating when the message needs to be sent to people more than the limit. This prompts users to consider other social networking platforms.</p><h4>Issue#5</h4><p>Can not edit sent Status</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VZhJc5VueOwh8CDmsYk4PA.jpeg" /></figure><p><strong>Severity: </strong>2</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>A user can not edit a sent status. In case a video is uploaded and needs to be edited, the user would need to re-upload the entire video to edit the status. Though a quick survey would indicate the severity, editing status definitely puts the user through the pain of deleting, re-uploading, and press a button to put the status again.</p><h4>Issue#6</h4><p>Different ways to pin a chat on different platforms</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/821/1*NveRBeVhtvSCw8ANIrKAWA.png" /><figcaption>(Image by <a href="https://www.howtogeek.com/333932/how-to-pin-whatsapp-conversations-to-the-top-of-the-chat-list/">HOWTOGEEK</a>)</figcaption></figure><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> Consistency and standards</p><p><strong>Description: </strong>An Android user can pin a chat by long press and then selecting the pin option in the app, while an iOS user would need to right swipe to pin a chat. The user has to go through the pain of searching how to do the same in case of switches to the other platform.</p><h4>Issue#7</h4><p>The caption(message) is not forwarded with the image</p><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> Consistency and standards, Match between system and the real world</p><p><strong>Description: </strong>The meaning of `forward` refers to send on to a further destination. Forwarding a message consisting of an image with a caption loses the caption automatically. A caption for an image becomes an integral part of it, hence this is inconsistent with the purpose of forwarding.</p><h4>Issue#8</h4><p>Audios(Voice Message) attached(that are recorded in place) are sent without confirmation, unlike the case in videos</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FzN7LllEhbfw%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DzN7LllEhbfw&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FzN7LllEhbfw%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/2c171cda8b57b15c2c4eb37f1658862b/href">https://medium.com/media/2c171cda8b57b15c2c4eb37f1658862b/href</a></iframe><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> User control and freedom, Consistency, and Standards</p><p><strong>Description: </strong>Whenever a user records audio, as soon as the tap is released the file is sent without any consent, while in the case when videos, due consent is asked for. This issue thus limits the user&#39;s freedom and is inconsistent with the behavior while uploading other forms of attachments.</p><h4>Issue#9</h4><p>The video seen(Read Receipt) not indicated in the sender’s view unlike as done with audios</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/0*8RF9RqJog2XFg0de.jpg" /><figcaption>(Read receipts when audios are played)</figcaption></figure><p><strong>Severity: </strong>2</p><p><strong>Heuristics Violated:</strong> Visibility of system status, Consistency, and Standards</p><p><strong>Description: </strong>If the receiver plays the video, unlike the case of audios, no seen confirmation is sent. This is inconsistent and limits the visibility of system status at the sender’s end.</p><h4>Issue#10</h4><p>Unable to delete messages for others after an hour</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/498/1*k-8PHQ-pBH1ycFkbEBcfPQ.png" /></figure><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> User control and freedom</p><p><strong>Description: </strong>A user can not delete a message for others after a certain period of time (an hour). This limits the user control over as sent messages that are no longer intended to appear in the chat can’t be deleted.</p><h4>Issue#11</h4><p>Unable to apply different privacy settings to different contacts</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zfwE1_NxV-vP9Gc3.png" /><figcaption>(Image by <a href="https://iphone-tricks.com/tutorial/296-whatsapp-last-seen-how-to-turn-it-off">iPhone Tricks</a>)</figcaption></figure><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>Whenever WhatsApp’s privacy is being changed, a user can not add different privacy settings for a different contact. An interaction that solves this mapping for each user will make users feel safer and create a personalized experience while using the application.</p><h4>Issue#12</h4><p>The appearance of “This message was deleted” in case a user deleted a message for everyone</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*deGrqIPQTD52MFy1.jpg" /></figure><p><strong>Severity: </strong>4</p><p><strong>Heuristics Violated:</strong> Match between system and the real world</p><p><strong>Description: </strong>If a user is deleting a message before the other person has read the message, the user intends that the message shall disappear, while it gets replaced with “This message was deleted”. Though it is debatable I feel that the intention of the user is to see the message disappear once they delete it. A relevant User Quote:</p><p>“- Oh god! Why tell the other person that ‘<a href="https://blog.whatsapp.com/10000635/Deleting-Messages-for-Everyone">This message was deleted</a>’ — if one is deleting it before the other person has read the message, it should simply disappear. It is utter foolishness, triggering another conversation asking ‘What did you delete?’ &gt;Why designers why?&lt;” ~ <a href="https://blog.usejournal.com/why-i-hate-the-design-of-whatsapp-b5e683f1328">Yuvraj</a></p><h4>Issue#13</h4><p>No `personalized` notification in case a user is tagged in a message</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/914/1*k1xDKqbhPMvd25eNJXCHwQ.png" /><figcaption>(Image by <a href="https://www.wikihow.com/Tag-Someone-on-WhatsApp-on-Android">WikiHow</a>)</figcaption></figure><p><strong>Severity: </strong>2</p><p><strong>Heuristics Violated:</strong> Match between system and the real world, Visibility of system status</p><p><strong>Description: </strong>As a student I see that in a group of 5 or more, assigning tasks, putting out reminders is often done via WhatsApp. Whenever a user tags others in a message, there is not personalized notification other than the group message notification that focuses the attention of the desired user to the message.</p><h4>Issue#14</h4><p>Unable to block a group</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/539/1*Ac25EZ0P9QPUHzcFyVXgqQ.png" /></figure><p><strong>Severity: </strong>2</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>While blocking users separately is possible, blocking groups, thereby decreasing from any other interactions in the group is not possible in WhatsApp. It gets really frustrating when somebody keeps adding the other even if once disinterest has been shown.</p><h4>Issue#15</h4><p>Unable to prioritize messages and status by close friends</p><p><strong>Severity: </strong>3</p><p><strong>Heuristics Violated:</strong> Flexibility and efficiency of use</p><p><strong>Description: </strong>With Instagram and Facebook offering stories from close friends, introducing close friends in WhatsApp where chat and status priority would be given to them is very much awaited by its’ many users.</p><h3>Did I miss something?</h3><p>WhatsApp is an organization with highly skilled individuals. I believe they have a lot of thought put in them by some of the greatest heads of the design industry. So I think every decision they took had a quite broad and deep perspective and me critiquing those decisions as a part of the five-day design task is prone to misjudgments and errors.</p><p>Nevertheless, I tried my best to avoid deviations and keep my findings close to the domain of the problem statement.</p><p>What do you think of this review? Let me know!</p><h3>References</h3><ul><li><a href="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&amp;arnumber=8860063">Do Online Reviews Reveal Mobile Application Usability and User Experience? The Case of WhatsApp - IEEE Conference Publication</a></li><li><a href="https://www.interaction-design.org/literature/article/how-to-conduct-a-heuristic-evaluation-for-usability-in-hci-and-information-visualization">How to Conduct a Heuristic Evaluation for Usability in HCI and Information Visualization</a></li><li><a href="https://www.nngroup.com/articles/ten-usability-heuristics/">10 Usability Heuristics for User Interface Design</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d1b65da18e7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/a-ux-centered-analysis-of-information-sharing-and-privacy-over-whatsapp-d1b65da18e7">Case study: A UX-centered analysis of information sharing and privacy over WhatsApp</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Case study: Wanna GRUBIIIT?]]></title>
            <link>https://medium.com/design-bootcamp/wanna-grubiiit-eb8ded2b0922?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb8ded2b0922</guid>
            <category><![CDATA[ux-research]]></category>
            <category><![CDATA[case-study]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[hci]]></category>
            <category><![CDATA[ux]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Tue, 13 Oct 2020 16:34:12 GMT</pubDate>
            <atom:updated>2020-11-08T19:48:05.057Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>It’s 3 AM, and it’s morning at IIIT-Delhi. With around 30 students in the canteen, overwhelming number of orders were placed. My friend had to wait for another 30 minutes to get his, to regret that he wasted his time and could’ve done something that’d help him pass the mid-semester examination. But I was like it won’t matter much, hence ignored it.</blockquote><p>The next semester was this course, Human-Computer Interaction taught by <a href="https://www.linkedin.com/in/graceden/">Dr. Grace Eden</a> that led us ideate into issues that we faced in our daily life at IIIT-Delhi and work on them. Where it led us? Stay tuned :D</p><figure><img alt="GRUBIIIT Cover — Think, click &amp; pick food of your choice. Save your time and say no to waiting in queues!" src="https://cdn-images-1.medium.com/max/1024/1*LQ9kkfgdLACIkPiWzNtkKw.png" /></figure><h3><strong>Introduction</strong></h3><p><strong>The Team</strong></p><figure><img alt="Team Photograph!" src="https://cdn-images-1.medium.com/max/1024/1*VArDnSPYLjoVIAZCKg6ZMQ.jpeg" /></figure><p>We had an amazing team where we had lots of fun, did some good work and applauded each other while criticizing to the max. extent! Members: [<a href="https://medium.com/u/b3eccd15196a">Aditya Sharma</a>, Kunal Anand, Mukul Tomar, Rakshita Anand, Rishi Raj Jain]</p><p><strong>The search</strong></p><p>We arrived to her with these set of ideas on the theme ‘HCI and Time’:</p><p>1: Relying on google maps at the need of the hour, consumes user time and wastage of resources.</p><p>2: Inability of current psychiatric condition tracking devices to study the patient overtime &amp; analyze the improvement in the conditions.</p><p>3: Procrastinating behavior and the inability of the user to make future decisions based on the growth, causing imbalance in one’s schedule.</p><p>[Find the <a href="https://docs.google.com/presentation/d/1MLbQguwVB9ys3nypnFCeGhCZp_8zNIenavnNti-wc0A/edit#slide=id.g6e3d1eb88e_0_275">detailed proposal</a>]</p><p>While the second idea convinced us of it’s innovativeness, feasibility was a concern. With the aim to learn the process, the discussion with the professor concluded with that these ideas do require lot of input to work on with. But here’s the interesting the part, we got an idea from the professor herself, she faced the difficulty to walk down the nearly whole campus to reach canteen. She proposed ‘What about a food app for IIIT-D?’ and here are the set of things that got me all in for it, immediately!</p><p><strong>Motivation and Inspiration</strong></p><p>• Well, nobody wants to stand in long queues to order food, right? As a IIIT-Dian, say you’ve a deadline at 12 PM and you are very hungry as you could not go to mess for your dinner. What do you do now!?</p><p>• Life’s easier. Order and get back to work, pick up food on the way.</p><p>• What if I’ve had a friend bring my order? That’d be so cool!</p><p><strong>Problem Statement</strong></p><p>“A customer trying to order food from food corners at IIIT-D, but the wastage of time caused due to standing in the long queues and waiting for the order to get prepared leaves him/her frustrated.”</p><p>We arrived at this problem statement after several iterations. Herein, we focused on <strong>Capturing the user and the need, </strong>after all problem statement identifies the gap between the current (problem) state and desired (goal) state of a process or product. [<a href="https://en.wikipedia.org/wiki/Problem_statement">Wikipedia</a>]</p><h3>Competitive Analysis</h3><p>“A competitive analysis is a way to collect and compare data about products(and companies) in the marketplace. This method is often used to highlight the strengths and the weaknesses of products in order to make more informed decisions about your product strategy.” — (Dalrymple, 2018)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*vDQgRkqAZea6FpHQCZX16A.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*R5H-r_z-jxTzVEJhsSwjMQ.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*BYBv6YYlVKTYLFwyBETimg.png" /><figcaption>(Competitors: Zomato, Swiggy &amp; Dominos)</figcaption></figure><p>Lookup our detailed Competitive Analysis at: <a href="https://drive.google.com/file/d/19j7wgYx1E5A7qfIIXDqxaepcb8Xa3F6r/view?usp=sharing">https://drive.google.com/file/d/19j7wgYx1E5A7qfIIXDqxaepcb8Xa3F6r/view?usp=sharing</a></p><p><strong>Summary</strong></p><p>While Domino’s app definitely has a smoother ordering process than Swiggy and Zomato, they’ve lacked at addressing quick search. Swiggy and Zomato posed issues such as cluttered user-interface and non-removal of closed restaurants.</p><p>Though the competitive analysis lead us to the features meant to address some of the issues, it also limits our ability to think. This posed a need for conducting User Research as to gain insights from the user themselves.</p><h3>User Research</h3><p>User research is conducted so as to understand users’ characteristics, aims, and behaviors towards achieving these aims. Its purpose is to produce designs that improve their working practices and lives. [<a href="https://www.interaction-design.org/literature/topics/user-research">Ref.</a>]</p><p>We proceeded with the Qualitative Research where we develop a deep understanding of <strong>why</strong> users behave the way they do. We conducted both, semi-structured and contextual interviews to gather such understanding.</p><p><strong>Interview Process</strong></p><p>We conducted 3 semi-structured and 2 contextual interviews, from B.Tech to Ph.D we tried to have at least one interview to bring diversity to the set within the time and course limitations.</p><p><strong>Relevant User Quotes</strong></p><p>“People who ordered for the same item after them, if reached to the counter before, get to receive it, as there is no proper mechanism for delivering the order as the workers there shout and call the order that is to be served.”</p><p>“The quality of food that I order from outside the campus is better than at food corners at IIIT-Delhi, and yes the food comes directly to the hostel.”</p><p><strong>Summary</strong></p><p>Me, specially was motivated after conducting interviews as it confirmed us of our suspicion! What else did I need? The problem we were looking for exists :D Talking to variety of people helped me understand what are their frustrations, needs, goals and expectations from the system.</p><h3>Affinity Mapping</h3><p>“The Affinity Diagram is a method which can help you gather large amounts of data and organize them into groups or themes based on their relationships. The affinity process is great for grouping data gathered during research or ideas generated during Brainstorms.” — <a href="https://www.interaction-design.org/literature/article/affinity-diagrams-learn-how-to-cluster-and-bundle-ideas-and-facts">Interaction Design</a></p><figure><img alt="GRUBIIIT — Affinity Mapping" src="https://cdn-images-1.medium.com/max/1024/1*PnKPt9Eh8QdANMRDvK1uVA.png" /></figure><p>This is our final affinity mapping. We conducted three iterations as to be precise and cover every aspect that seemed feasible within the limitations.</p><h3>Persona — Scenarios</h3><p>A scenario refers to the “descriptions of natural, constructed or imagined contexts for user-product interactions.” - (Sabine Madsen and Lene Nielsen, 2010)</p><p><strong>Who we chose and why?</strong></p><p>We chose three personas, a Ph.D. in biotechnology at IIIT-Delhi, a cash manager at sunrise food corner and a B.Tech student. Apart from their different work life and origins, they possessed different interests, goals, frustrations and the expectations from the food life at IIIT-Delhi. This was a key factor in deciding the selection of the personas, where each had a different insight to offer for the same goal that we wanted to achieve. This proved to be fruitful as we got to know of the issues such as communication being a barrier in placing the order, the need to be physically present at each food corner to know the menu, digital payment being restricted to Paytm, absence of feedback at some food corners.</p><figure><img alt="Persona 1 — Anuj Sharma" src="https://cdn-images-1.medium.com/max/1024/1*49beEYKVUHznzR5J-i7erw.png" /></figure><figure><img alt="Persona 2 — Mukesh Kumar" src="https://cdn-images-1.medium.com/max/1024/1*cM208z1_BXT9hUe8rhcvuQ.png" /></figure><figure><img alt="Persona 3 — Sagar" src="https://cdn-images-1.medium.com/max/1024/1*Ry1gO9jr1kolAh7UJS8bWQ.png" /></figure><h3>Storyboards</h3><p>“Storyboards are illustrations that represent shots that ultimately represent a story. Basically, it’s a sequential art, where images are arrayed together to visualise the story.” — Nick Babich, 2017</p><figure><img alt="GRUBIIIT — Storyboard" src="https://cdn-images-1.medium.com/max/1024/0*wKKyb8Tf8OFLleR9" /></figure><p><strong>Why did we choose to develop this story?</strong></p><p>In our story, the character was a B.Tech student residing in a hostel. The scene is set in his room, with his roommate right there with him. The plot is such that the student has a deadline at a particular time but then he’s perplexed how to get food from the canteen without going to it, so as to work for the deadline meanwhile. His friend suggests him to use GRUBIIIT to place the order for pickup. Twenty minutes after ordering he receives the notification that his order is ready for pickup. This story offers the reader:</p><ol><li>Work-life of the character at IIIT-Delhi.</li><li>The constraints that the character is facing.</li><li>The confusion and hesitation to sacrifice one thing for the other, corresponding to the current situation.</li><li>How easily the awareness of the application will grow within the hostel.</li><li>Notification system integrated into the application.</li><li>Bring out the need for such an application.</li></ol><p><strong>My Learnings</strong></p><p>While brainstorming for the storyboard, each of us came across many different ideas. While the current scenario covers the problem and presents the need for such an app, I had an input that ‘ what if we show that the other person is picking up the order?’ That’d convey that the application also provides the feature for someone to rely on others while they are busy doing something. From this activity I learnt how a designer can present a case to someone who does not possess any knowledge about an issue, delivering the message effectively using words and visual elements.</p><h3>Information Architecture</h3><p>An unplanned way to organize content in the application takes us back to square one, users might feel frustrated and might not give even it a second chance and uninstall the app! This is why it is important to have a planned, diagrammatic workflow of the application. Have a look at our broad architecture that gives an overview of the application:</p><figure><img alt="IA — GRUBIIIT — Broad &amp; Shallow" src="https://cdn-images-1.medium.com/max/1024/1*dycaubD6ShwxYVUVtnuWnA.png" /><figcaption>(Overview: Broad and Shallow Information Architecture)</figcaption></figure><p><strong>My Learnings</strong></p><p>At each and every stage of creating the diagram I felt more conscious of what I am doing, and how things are going to work out. The workflow on completion, showed a definitive manner in which a particular user scenario be handled. Also from the feedback of our professor, we came to know that:</p><ol><li>‘Filters’ don’t belong to an information architecture,</li><li>Enough focus on USP shall be visible.</li></ol><h3>Prototyping</h3><p>A prototype is a draft version of a product that allows you to explore your ideas and show the intention behind a feature or the overall design concept to users before investing time and money into development. A prototype can be anything from paper drawings (low-fidelity) to something that allows click-through of a few pieces of content to a fully functioning site (high-fidelity) — <a href="https://www.usability.gov/how-to-and-tools/methods/prototyping.html">Usability Gov, 2019</a>.</p><p><strong>Low-Fidelity Prototype [Individually done by members of YouEx]</strong></p><p>Wireframes need not contain fancy UI, they can but it’s not necessary. It should just reflect the skeleton of the UI.</p><figure><img alt="Low Fidelity!" src="https://cdn-images-1.medium.com/max/1024/1*Vz6QrkjvQ2VcD7OIe0vhDw.png" /><figcaption>(Certainly I need to level up my drawing skills)</figcaption></figure><p>While I made the low fidelity sketches I was quick with drawing out whatever just came to my mind on paper to represent/communicate the rough idea of what I had thought as the most important thing should be available to the user in the thought scenario. The process that I followed in drawing them was by imagining a scenario where a new user wants to use GRUBIIIT and wants to place order after applying filters to the food corners.</p><p><strong>Mid-Fidelity Prototype</strong></p><p>While doing the medium fidelity prototyping, I thought of it as a little more detailed version of the low fidelity one. It’s like adding a life to a skeleton.</p><figure><img alt="Medium Fidelity!" src="https://cdn-images-1.medium.com/max/1024/1*_OwRmDFWZUk69kbGeU5Y7g.png" /></figure><p><strong>High-Fidelity Prototype</strong></p><p>Again, I was honored to have the opportunity to build the wireframes for the group and as soon as I completed it, I surely knew I learnt a lot. While doing high fidelity prototyping, apart from adding color, images and actual text to the placeholder, the major time was spent in iterating on how users would operate on such a prototype. Basically in high fidelity we sort of evaluated on complex-interactions. We iterated over issues in our low and mid-fi and introduced a new features named ‘GRUBfeed’ &amp; ‘Pickup Buddy’ in the final deliverable.</p><figure><img alt="High Fidelity!" src="https://cdn-images-1.medium.com/max/1024/1*DMNb2Kl5anc6U0qE0EhDxA.png" /></figure><h3>Prototyping Evaluation [Guerrilla Testing]</h3><p>“Guerrilla testing is a means of gathering user feedback by taking a design or prototype into the public domain and asking passersby for their thoughts. Due to its simplicity, new ideas can be tested quickly and at a low cost, making it a valuable UX testing method.” — Elizabeth Chesters, 2017</p><p><strong>Evaluation Process</strong></p><p>We selected a scenario which the user would be asked to execute. One sat, observed and wrote down the difficulties users were facing and any additional comments they had.</p><p><strong>The Scenario</strong></p><p>“So you have to use our app and see what others are posting on social forums. Place the food order from the post directly that you find interesting and request if any user can pick it up for you.” — described general actions, where we intended the user to use ‘GRUBfeed’ in the application to proceed with the given task.</p><p><strong>Who did we select to evaluate our prototype and why?</strong></p><p>We had student hostellers and day scholars from different years to evaluate our high fidelity prototype. We selected such an audience so as to obtain a wide range of insights and critique according to one’s experience.</p><p><strong>Relevant User Quotes</strong></p><p>“GRUB feed should be made GRUB chat because it looks more like a chat rather than a feed.”</p><p>“Scenario was easy to execute except the pickup buddy button at the order confirmation screen should be made look more like a button and I wasn’t aware of what it was also so some info to tell the users what it is would be helpful.”</p><p>“First of all, I think GRUBfeed is pushed to the corner, GRUBfeed should be in the centre. You should change the name, and the cart should be in the corner. Usage of tags on items listed in the GRUBfeed seems helpful in narrowing my search. PickUp Buddy seems a very promising feature.”</p><p><strong>My Learnings</strong></p><p>Yes do think, but from the users mind only! — Even after going through such an extensive while designing the high fidelity, I can’t believe we missed many things. People were able to point out things that I couldn’t even imagine.</p><h3><strong>Prototyping Revision</strong></h3><p>It is the user we are designing for! After documenting the user feedback and preparing the summary of the issues and recommendations by the user, I would think deeper into those points and look for certain practices that might help in incorporating those changes effectively. One of the points that I recall was “GRUBfeed is pushed to the corner, bring it to the middle or center for drawing user’s attention” , will require a thought as the dynamics of the app would then change. This is an important issue, as GRUBfeed might just be the selling point of our app.</p><figure><img alt="Prototype Revision — Feature 1" src="https://cdn-images-1.medium.com/max/812/1*Qjjw_XL1NUc4yw8yeWw_Vg.png" /><figcaption>(Revision of GRUBfeed Page with the relevant User Quotes)</figcaption></figure><figure><img alt="Prototype Revision — Feature 2" src="https://cdn-images-1.medium.com/max/792/1*zUtC862ho56fSyh-iaXeVA.png" /><figcaption>(Revision of Order Confirmation Page with the relevant User Quotes)</figcaption></figure><h3><strong>Summary</strong></h3><figure><img alt="UX &amp; UI" src="https://cdn-images-1.medium.com/max/1024/0*YuaWMFPQoIYNhqa-.jpeg" /><figcaption><a href="https://miro.medium.com/max/1050/1*hp-yfKsmzsj711iLbM8eEw.jpeg">https://miro.medium.com/max/1050/1*hp-yfKsmzsj711iLbM8eEw.jpeg</a></figcaption></figure><p>While working on this, I had interactions lot of interactions with colleagues both formal and informal, read some of the articles as referred by our professor and practiced the complete design process. I wouldn’t have been writing all of this if I had not documented my whole work for the final submission, so that’s how documentation is. Thanks to <a href="https://www.linkedin.com/in/graceden/">Dr. Grace Eden</a>, <a href="https://www.linkedin.com/in/arpit-bhatia/">Arpit Bhatia</a> and <a href="https://www.linkedin.com/in/anupriya-tuli/">Anupriya Tuli</a> for their guidance, regular feedback and motivation!</p><h3>References</h3><ol><li>Mills, D. wiki.c2.com(2014) Problem Domain</li><li>Dalryple, B., medium.com (2018) Competitive Analysis — User Research</li></ol><p>3. Persona-Scenarios — <a href="https://classroom.google.com/u/1/c/NDc3NTM4MzgzOTFa/m/NTI">https://classroom.google.com/u/1/c/NDc3NTM4MzgzOTFa/m/NTI</a> yMDIwNTU1NjJa/details</p><p>4. Babich. N., uxplanet.org (2017) Storyboarding in UX Design</p><p>5. Studio. B, uxplanet.org (2017) Information Architecture. Basics for Designers.</p><p>6. usability.gov (2019) Prototyping</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb8ded2b0922" width="1" height="1" alt=""><hr><p><a href="https://medium.com/design-bootcamp/wanna-grubiiit-eb8ded2b0922">Case study: Wanna GRUBIIIT?</a> was originally published in <a href="https://medium.com/design-bootcamp">Bootcamp</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Mobile, Web Apps? Get started with React Native. API? Use Apollo GraphQL.]]></title>
            <link>https://medium.com/@rishi-raj-jain/mobile-web-apps-get-started-with-react-native-api-use-apollo-graphql-71ab384a8be7?source=rss-cf48f7f14d64------2</link>
            <guid isPermaLink="false">https://medium.com/p/71ab384a8be7</guid>
            <category><![CDATA[apollo-graphql]]></category>
            <category><![CDATA[expo]]></category>
            <category><![CDATA[react-native]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[graphql]]></category>
            <dc:creator><![CDATA[Rishi Raj Jain]]></dc:creator>
            <pubDate>Mon, 20 Jul 2020 15:39:48 GMT</pubDate>
            <atom:updated>2021-02-19T15:31:55.150Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*JGQQHoDXsvAjN0KcMrkbwg.png" /><figcaption>United: GraphQL, Apollo and React Native (by <a href="https://www.freecodecamp.org/news/content/images/size/w2000/2020/04/cover-3.png">Freecodecamp</a>)</figcaption></figure><p>It’s more than a year I’ve been doing web development and when I came across React, React Native, GraphQL API, and Apollo GraphQL, I was in complete awe to what they could do and how much time would they save! Whoa, hold on hold on, don’t get scared by the beauty of them already. This blog will keep you informed on what is React Native, how GraphQL is better than Rest API, and what purpose does Apollo GraphQL (Client) serves.</p><h3><strong>What are native applications?</strong></h3><blockquote>Native applications are compiled into the machine language of the running CPU. For example, Windows and Mac executable apps are in x86 machine language, while mobile apps are ARM based (see <a href="https://www.pcmag.com/encyclopedia/term/x86">x86</a> and <a href="https://www.pcmag.com/encyclopedia/term/arm">ARM</a>). (Source: <a href="https://www.pcmag.com/encyclopedia/term/native-application#:~:text=An%20executable%20program%20coded%20in,(see%20x86%20and%20ARM).">PCMAG</a>)</blockquote><h3>React Native — A framework for building native apps using react.</h3><p>As widely advertised, react-native combines native development with React, a JavaScript library for building user interfaces. It enables the creation of purely native apps, that maps to the platform’s UI elements. So be it Web, iOS, or Android, code once and enjoy! Heard of <a href="https://instagram-engineering.com/react-native-at-instagram-dd828a9a90c7#.3h4wir4zr">Instagram</a>? React Native is currently being used on Instagram, Tesla, <a href="https://eng.uber.com/ubereats-react-native/">Uber Eats</a>, <a href="https://blog.discord.com/how-discord-achieves-native-ios-performance-with-react-native-390c84dcd502">Discord</a>, and famous website builder <a href="https://www.wix.com/mobile/wix-app">WiX</a>.</p><p>An advantage of using React Native is the option to utilize <a href="https://expo.io/"><strong>Expo</strong></a>. Expo’s tools and services such as:</p><ol><li><strong>Templates: </strong>Lazy? Setup React native <a href="https://github.com/necolas/react-native-web">(with web)</a> &amp; navigation with `tabs` template.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6sb6iN39HrVJtU3-nm8lkQ.png" /><figcaption>Available templates with Expo CLI</figcaption></figure><ol><li><strong>Social Authentication: </strong>Enable log in with the Facebook and Google native SDKs, or support any browser-based auth flow.</li><li><strong>Push Notifications:</strong> Send and receive cross-platform push notifications with a few lines of JavaScript.</li><li><strong>Icons:</strong> Set of icons on top of FontAwesome &amp; Ionicons.</li><li><strong>Native Graphics: </strong>Expo includes a WebGL-compatible API called EXGL.</li></ol><p>Cool, right? Try it in the browser: <a href="https://snack.expo.io/">https://snack.expo.io/</a></p><p>Learn and explore more on <a href="https://reactnative.dev/docs/tutorial">React Native Tutorial</a> &amp; <a href="https://docs.expo.io/">Expo Documentation</a></p><h3><strong>GraphQL is here! Time to REST is over.</strong></h3><blockquote>GraphQL is a query language for an API, and a server-side runtime for executing queries by using a type system one’s define for the data. GraphQL isn’t tied to any specific database or storage engine and is instead backed by your existing code and data.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fZXXf587pg-OrIc-bPRZDA.png" /><figcaption>Image from <a href="https://graphql.org/">GraphQL</a></figcaption></figure><p><strong>Major Advantages of GraphQL over Rest API:</strong></p><ol><li>With a REST API, one would typically gather the data by accessing multiple endpoints, say /shops/&lt;id&gt; to fetch the initial data and then /shops/&lt;id&gt;/items to get all the items that would be available for the particular shop, in GraphQL a simple query to a <strong>single </strong>endpoint to the GraphQL server is sent which will return the required JSON.</li><li>No over fetching: Overfetching means that a client downloads more information than is actually required in the app. While one may hit /shops with the sole purpose to request the names of the shops, obtaining the same with rest API might also give things other than those required (say shop images as well in the JSON). With GraphQL, this kind of mutation is a cakewalk.</li><li>Easy resolution to under fetching: Underfetching generally means that a specific endpoint doesn’t provide enough of the required information. This can escalate to a situation where a client needs to first download a list of elements but then needs to make one additional request per element to fetch the required data. Rather than hitting multiple requests, GraphQL enables you to get this all of this in a single request.</li></ol><p>Learn more about GraphQL and how to use it to <a href="https://www.howtographql.com/">@howtograhql</a></p><h3><strong>Apollo GraphQL (Client)</strong></h3><blockquote><a href="https://www.apollographql.com/docs/react/">Simplify app development by combining APIs, databases, and microservices into a single data graph that you can query with GraphQL.</a></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/988/0*gMn_0g_fLlzDimPv.png" /><figcaption>Modern Architecture Management by <a href="https://www.apollographql.com/">Apollo GraphQL</a></figcaption></figure><p>Apollo is the industry-standard GraphQL implementation, providing the data graph layer that connects modern apps to the cloud.</p><blockquote><a href="https://www.apollographql.com/docs/react/"><strong>Apollo Client</strong> is a comprehensive <strong>state management library</strong> for JavaScript that enables you to manage both local and remote data with GraphQL.</a></blockquote><p>Features of Apollo Client:</p><ul><li><strong>Declarative data fetching:</strong> No need to manually track loading states, just write queries and let it handle the rest.</li><li><strong>Caching:</strong> Easier local-only modifications via querying the cache directly.</li><li><strong>Local State:</strong> Manage both local and remotely fetched state with a single API.</li><li><strong>Designed for modern React:</strong> Take advantage of the latest React features, such as hooks.</li><li><strong>Universally compatible:</strong> Use any build setup and any GraphQL API</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7uKF1HODRR0DNkuOJeM5dg.png" /><figcaption>Create Apollo App from (<a href="https://apolloapp.org/">apolloapp.org</a>)</figcaption></figure><p>Searching the web for a starter kit? I doubt you’ll find something better than <a href="https://apolloapp.org/">Create Apollo App</a>, that contains the starters of a web app, or native mobile app, even a universal application and lot more on server enabled apps.</p><p>I have also made a React Native (with react-native-web), Expo, Apollo Client &amp; GraphQL starter. Find it here at <a href="https://github.com/digitalPlayer1125/React-Native--With-Web--ApolloGraphQL-starter/">React-Native — With-Web — ApolloGraphQL-starter</a></p><p>That’s what I have for now. Hope this helps!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=71ab384a8be7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>