Tech stacks don't matter. Use what you know and build something first | Hacker News Reader

Tech stacks don't matter. Use what you know and build something first

4 pointsjakemanger5 comments

When I first started building websites and apps, I searched almost every day for whether I was using the most optimal tech stack.

It wasn't until I got used to building products in one, that I realised learning the newest "blazingly fast" framework or switching to some new database doesn't really matter if you have no users.

What matters is whether you can build something quick to test your idea. After learning that, I built my last project in a month and it surpassed $1000 in revenue in its first two weeks!

What am I doing now? Using the same tech stack to build my next idea.

View on HN

I agree. This is why, while I love knowing and messing around with other languages, I generally stick with Python unless I have a demonstrably good reason not to. (For web stuff specifically, Django is still my go-to.)

One good reason not to: I wanted to build something that cross-compiled to a single binary which could run on all major platforms, and benefited from a particularly heavy in-memory data structure so as not to burn through people's hard drives unnecessarily (though not much else). In that case I reached for Go, and ended up with https://github.com/hiAndrewQuinn/tsk powered by a big fat trie under the hood. Obviously tries aren't exactly groundbreaking technology, but Python's memory footprint would have made me uncomfortable, even if I did figure out how to ship it as a py2exe.

It's a bit of an over-generalization, though. Tech stack matters more and more, the larger a project gets, and the more people join it. Some languages and frameworks lend themselves better at long-term maintainability, as the code base gets larger. Other put so many foot-guns at your disposal, that you need a lot of discipline to keep things manageable. And, of course, once you start working together with others, the choice of language and framework becomes much more consequential, as it will limit who can join the project, how quickly they can become productive once they join, etc.

I promise not to knock it, but what was that stack? Thanks for sharing!

haha I'm vulnerable here. Settled on Nextjs, React, Supabase and Tailwindcss for most things and I can build things quite quick with it now. If I need a local application (e.g. for Mac, Windows or Linux), I reuse the same code with Tauri.

But look at people like Pieter Levels https://levels.io/how-i-build-my-minimum-viable-products/

He uses PHP because he knew it and could build things quick with it (in a single file in his case).

Reply on news.ycombinator.com