Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Doesn't seem to be a popular opinion here, but I think FE can be ok. A typical React application feels easier to reason about than HTMX-based code that I looked at. I never had a hard time picking up an app that used React/Preact + Tailwind + some simple state management library like Preact Signals.

Maybe never having had to deal with GraphQL or Server Components helped.

Tooling, OTOH, does feel like it gets insane from time to time. I don't know why we needed such a painful migration to go from CommonJS to ES Modules in Node.js. They always just 'worked' in Bun, and even Node.js has a flag these days to allow you to use both of them together. ESLint changed their configuration format and instantly made every existent tutorial/configuration available out of date. I fail to see how anyone benefited from that.

It's not as popular these days, but I appreciate how Meteor always made an effort to be backwards compatible to the extend it was possible. The only __major__ migration was going from Fibers to Promises - and I don't blame them for picking the wrong 'horse' 13 years ago.

But yeah I feel very comfortable doing FE these days. Meteor for bigger fullstack apps, or just Vite for a simple FE, and things just work. TypeScript tooling is better than ever. AI helps automating lots of the simpler but tiring tasks like generating Zod schemas.




I also want to add: the new hooks introduced by React 19 seem like a complexity creep. But I haven't found any need to use them yet.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: