It’s unfortunate but the component library situation is really Svelte’s weak point currently.
At Monitoro[0] we bit the bullet and implemented the vast majority of our components from scratch. Apart from the obvious time to develop and test, and the trailing bugs that are hard to solve for small closed source projects, the experience wasn’t that bad.
Ultimately what helped us the most is writing our components using a state machine-like pattern, mixed with TailwindCSS to make styling easier.
In our case, the effort was worth it as we anyway needed several super specific components, and now that we’re over the hill we have complete control on our UX.
(Also having built component libraries/design systems before in different UI frameworks, doing it in Svelte was one of the best experiences so far)
There are some efforts in the Svelte ecosystem but they’re small and do not have much firepower behind, thus limited or of relatively low quality.
Word, thanks! I think it's interesting because Svelte community support is pretty big. The repo's consistently been on Github's Trending repos for the past 2.5 years and it has an impressive number of stars
> been on Github's Trending repos for the past 2.5 years and it has an impressive number of stars
Wow, must mean the project is real important. I agree that svelte is interesting, but focusing on metrics and popularity is the wrong approach to new technology.
You can't judge a technology by its popularity. But I was judging the ecosystem by its popularity and those things are very highly correlated. More people bookmarking a project mean in theory it should have better 3rd party library support.
You're judging the ecosystem based on the stars/trendiness of Sveltes repository? That feels even more wrong and you're doing a lot of extrapolation here.
At Monitoro[0] we bit the bullet and implemented the vast majority of our components from scratch. Apart from the obvious time to develop and test, and the trailing bugs that are hard to solve for small closed source projects, the experience wasn’t that bad.
Ultimately what helped us the most is writing our components using a state machine-like pattern, mixed with TailwindCSS to make styling easier.
In our case, the effort was worth it as we anyway needed several super specific components, and now that we’re over the hill we have complete control on our UX.
(Also having built component libraries/design systems before in different UI frameworks, doing it in Svelte was one of the best experiences so far)
There are some efforts in the Svelte ecosystem but they’re small and do not have much firepower behind, thus limited or of relatively low quality.
[0]: https://www.monitoro.xyz