@voltstimesamps

Devlog #7: Results Page
Yay, the results page is done! Took maybe 30 mins to write the rendering code, 15 to wire everything, and another hour and a half to troubleshoot everything. Soooo many issues. Had to deal with CORS, Prisma migrations, database schema issues, discrepencies in .env and .gitignore files between my two workspaces, etc. After many-a-struggle, it’s done! I’m gonna do a system description page so you can click on each system in the results page, do a design pass, and then I think I can ship and deploy. Excited tho!
Thanks for reading and happy coding!

Devlog #6: Questionnaire
All of the question slides are done, working as I think they should, and flow into each other. Main time-suck for this part was getting all of the types containing the values for the buttons to correspond with parameters for the computer-finding function. Video has no cursor, sorry! I did realize while recording the video below that my longevity question doesn’‘t have the unit, so I’ll fix that. Right now, you have to select something before you can click continue, I’ll deal with how that visually appears with CSS. Excited to be reaching the point where I can ship, soonish.
Thanks for reading and happy coding!

Devlog #5: Questionnaire Frontend & React Struggles
I did some tidying up code-wise and fixed some bugs about null values, not gonna go deep into those.
More excitingly, THE FRONTEND!!
The questionnaire is coming together. I wrote the functions for sequencing its steps, handling answer inputs, and rendering each step. Everything is linked into page.tsx. Took a lot more time then it should, thanks to absolutely no knowledge of how React works or how to build a frontend, so learning as I go with the help of Claude and W3 among other resources. Super stoked though!
Thank you for reading and happy coding!

Devlog #4: Frontend Init & Backend Cleanup
Initialized Next.js frontend with some AI help explaining what I need and don’t need for a custom config. I realized that for the frontend, I would also want to base operating systems on types rather than versions, so I added an enum type and updated my data to follow that system. Lastly, I decided that I want to use app tiles where people can select what apps they use, and then there will be an algorithm that computes those selected apps into computer specs, so I created a dataset with those apps. Spreadsheet with the apps pictured below.
Thanks for reading and happy coding!

Devlog #3: More Unit Tests & Fixing Rec Logic
Alright. So I’ve finished the unit tests. It took me less time than I expected, albeit there was a lot of copy-and-pasting for creating new fake computer systems to test. A couple of issues arose.

1. Parts of the total score were exceeding their assigned values (xxxWeight), so I clamped it using Math.min.
2. After vramGb met the minimum, all GPUs would score the same. Fixed by multiplying the weight for GPU score inside the Math.min, taking the minimum of the calculated score and the assigned weight.

Super satisfying to see al of the tests past. Now, on to the frontend!
Thanks for reading and happy coding!

Devlog #2: CRUD Routes & Vitest Unit Tests
Been fighting with Vitest on and off for two hours now 😢 but I’ve finally got it. Built a test for my hardFilter function, which removes systems in my database if they do not meet the base requirements. Included them for almost every parameter in my system requirements profile. The most difficult part was making it work with my cursed nested tables (thank you, Prisma) that I have connecting my computer parts to the systems that they are in. It was long and mostly copy-and-pasting the same tables over and over again, but it’s done and done well (I think/hope/pray).
CRUD routes are made, not implemented. For now, all I can do is query via GET route through my dev script, but hoping to be able to use them to manage my database through a frontend eventually.
Thanks for reading and happy coding!

Hello everyone!
I’m working on a PC recommendation website. I was looking for a new laptop and realized that there wasn’t really a well-known platform for finding computers, so I’m trying to make one.
I’m building full stack, including my own database schema on a managed PostgreSQL instance. I’m using Prisma, Hono, and Neon for my backend, Next.js and Vercel for my frontend hosting, and either Railway or Render (haven’t decided yet because I’ve never used either) for backend hosting. Trying to keep it low(ish) budget. Coding in TypeScript.
Before starting Stardance, I had my database structure set up and was working on seeding the database.
Since I’ve started logging, I have:
-Finished the seed file
-Created a ranking, scoring, and filtering system for my recommendation engine
-Created REST API endpoints(GET for querying the database and a POST route for the recommendation engine)
-Mounted everything through Hono
-Added Next.js routes for testing and implementation
Any advice or questions from the community are appreciated. I have coding experience but not web development experience, so most of my info is coming from friends, online, communities, and AI (fact-checked and responsibly). Thank you and happy coding!
Screenshots below: a list of all of the systems in my database, a query to my recommendation engine for a laptop with at least 16 GB of ram, and a list of all of the manufacturers in my database, all built through Hono and Next.js.