🔭Telescope: Explore Github Stars

Jeremy, a 38 year-old developer that looks 80

What is Telescope?

Telescope is a NextJS demo app that allows a user to search the stars of any GitHub user. It was built in a short 1-week timeline using the following technologies:

  • NextJS (React)
  • GraphQL
  • Tailwind CSS
  • Jest

You can find the code on GitHub

Project Requirements

The project needed to fit the following requirements:

  • Use a publicly available API
  • Have at least two views: an overview and a details page
  • Be thoroughly tested and accessible

Pretty easy right? I thought it would be, but there were definite challenges - mainly corresponding with the choices I made as outlined below (KISS).

Design Choices

I wanted to use building Telescope as a learning experience, and opted to try some technologies I have less experience with.

Create React App vs NextJS

I struggled with a decesion between using Create React App or NextJS. The project requirements made this seem like an easy choice for CRA (single page app). But like a moth to a flame I was drawn to that new hotness. Yes, I'm talking about NestJS 13, with the new (still in beta) app directory layout. This did simplify building SPA style routing, but introduced a host of complexity to testing and general overhead. I also wanted to see if I could improve accessibility, SEO and performance with NextJS.

JavaScript vs TypeScript

Admittedly (and as I'm sure you can tell reading the code) I do not have a ton of experience using TypeScript. I last used TypeScript professionally over 2 years ago, and was pleasantly surprised at how much the tooling has improved. I was able to get up and running with TypeScript in NextJS in a matter of minutes, and the IDE support was great. However, my lack of experience here led to some frustration with the type system, and I found myself spending more time than I would have liked trying to get those red squigleeze to go away. I also found myself using any in more places than I would have liked. Utlimately though I can see the benefits of investing time up front using TypeScript. It seems to be a very useful tool and I will continue learning it and using it in my projects.

PHPStorm (Jetbrains) vs VSCode

I typically use PHPStorm in my day-to-day, but wanted to give VSCode a chance. It seems like most front-end engineers use VSCode, and it looked like the TypeScript, GraphQL and NextJS support were easier to set up than in PHPStorm. I was very happy with the experience in VSCode but do need to either configure the hotkeys to be the same as PHPStorm or train my muscle memory some more.

REST vs GraphQL

Similarly to the choice between JavaScript and TypeScript, I use REST APIs daily, while I haven't worked with GraphQL professionally in over 2 years. I again was pleasantly surprise by how the tooling has improved in that time. The GitHub GraphQL documentation was great and I was able to grab just the data I needed with very little effort (although I'm sure I could clean up the queries a bit more).

MacOS vs Windows

I typically do my professional and personal development on a Mac, but since I was treating this project as an exercise in masochism, I might as go all the way right? Thank goodness for Windows Subsystem for Linux (WSL) it's integration with VSCode on Windows 11. It might actually be easier to get a project up and running on a Windows machine now than a Mac.

Cypress vs Playwright

I've used Cypress for a few years now and have been very happy with it. I was excited to try Playwright as an alternative. It was super simple to set up, and has a built in test generator. I may need to implement Playwright in my day-to-day work.

Vue.js vs React

Well, this one really isn't a choice: it was in the requirements. I just wanted to mention that professionally I've been using Vue.js exclusively for some time. While I have done some personal projects in React over the last two years, It was a bit challenging to break my mind out of the "Vue.js way" of doing things.

Potential Future Improvements

There are a number of items I had as "stretch goals" that I wasn't able to accomplish in the short timeline.

  • Add a graph of the different repository languages a user has starred. We're already pulling this information down.
  • Add Supabase integration - allow users to log in and star repositories directly.
  • Night theme - not having a starry background feels like a huge miss.
  • UI/UX Improvements - There are places where the loading isn't tight and interactivity isn't highlighted.
  • Caching = I tried getting SWR up and running with GraphQL but hit a roadblock and ran out of time.

What this isn't

Perfect. I clearly didn't get to all I wanted to. The test coverage isn't all there. While the app should be accessible the performance didn't pan out. There are still a few red squiggles in the codebase, any is sprinkled throughout and I'm still not 100% sure how to properly use generics. However, I do think this was a successful exercise and learning experience, and I'm excited to continue to work on and improve the project in the future