(We have no external webapps because my company, funny enough, builds a JavaScript framework!)
(We have no external webapps because my company, funny enough, builds a JavaScript framework!)
Svelte was first released in 2016 and has become bigger than its original creator, Rich Harris. It has quite a thriving ecosystem and a big community. The tooling around it is more mature than React's has ever been.
> surface is so small
I actually find the opposite to be true. React still has a relatively small API surface area. One-way data flow makes it easier to understand. However, a smaller surface means arguably less expressivity, so I think Svelte's bigger learning curve is worth it. You can learn about 90% of React in a weekend and be productive. You can learn about 20% of Svelte in a weekend and also be somewhat productive, but you need to learn about 60-70% more to use Svelte as it's intended, rather than as a React alternative.
You only need to learn how to put {varName} in svelte.
.. is the “Nobody Got Fired For Buying (IBM|Microsoft|Cisco)” for the 2020s
Like NextAuth.js -> Auth.js React Query -> Tanstack Query
So at the very least, I think we are probably headed towards a plurality where it is practical to use Vue, React, Svelte or Solid though I do hope Svelte “wins” in the end.
I have to say, I still don’t think it’s to the point where it can be foisted on a team of mediocre developers in some kind of enterprise setting. Some of the libraries that exist for SvelteKit that I consider super critical are very small and maintained by one or two people.
I recently took on some freelance work for the first time in like 6 years, just to work on another large SK codebase and saw a lot of the same problems I encountered approached differently. I think there’s some key things that need to make it into core still.
Long-time Java dev by background, used more frameworks than I could possibly list. I won't touch React with a ten foot pole.
SvelteKit is the first web framework I ever used that I enjoy working with.