I still prefer svelte but it's less mature and universally-known. React is still a pretty good choice if you need something that will more or less work and that anyone can write.
I know svelte/sveltekit and would want to contribute to svelte apps (a good reminder that I should)
But there are some projects that I really really want to contribute to / heck even port to sveltekit like cinny and hedgedoc and the likes and so I almost feel pressured by the system to learn react so that I can finally feel confident enough to understand their repositories as they scare me right now...
Cinny:https://github.com/cinnyapp/cinny Hedgedoc:https://github.com/hedgedoc/hedgedoc
I don’t even want to talk to somebody that defines themselves as a “React” developer. Let alone work with them.
If you know any of these frameworks, and the basics of web development you can work in any of them.
So the hiring part is not an excuse for this.
Isn't it mainly about playing nice with crawlers? SEO and the like?
(that was my understanding but I'm a backend dev).
Honestly except for the marketing page and blogs and stuff, most apps are fine without server rendering. In fact I'd say many that avoid server rendering actually feel better simply because next.js makes it really easy to screw up your performance doing it.
I see this happening in finance data sites. Say a page about Apple, has stock price, etc. when logged in, same stuff but 10x data so they layout and everything is different.
Lots of newcomers are struggling and not understanding what are the options and which approach is best for their case.
Business people don’t help as they rightfully don’t care. But they want „do everything” - „pay once” approach so people bolt on static pages ona apps or other way around.
Example: This logistics SPA I was building I realized could just be single pages with some data for most of the stuff (tracking, inventory, etc...) but for admins they wanted a whole dashboard. This was a conditional on some value of the stored session user. So it ended up being kinda a website for parts of it and an SPA admin panel if the user conditionally matched some privileges. Probably should have been separate stacks but they used the same data so early on they made it the same Next app.
I don't think the whole website vs app thing is always as simple as static blog pages vs full fledged JS-heavy app. There is a spectrum and overlap even within a single "application" because of various requirements.
Your last sentence is the most accurate. I don't think its primarily ignorance, its just trying to meet all the requirements while retaining some level of organization in the codebase.
That being said, I'm waiting in the back stage, like many other folks, for tanstack to get production-ready, because of the all the weird crap being pulled by Vercel on NextJS.
What are the reasons for not doing SSR?
My default is a small page that client then fetches any additional data it needs. If its long load time skeleton UI it. I also have not seen the SEO benefits at all.
So again _why would I_ unless I needed to do stuff on the server to make the client bundle, which I don't.
A lot of these YC companies doing this could literally just be using a fetch because their backend is dead simple REST.