We need new website tech
Our current website is built with Jekyll. Keeping that maintained locally is a pain, especially as a not-a-ruby-dev.
I suggest we switch to something else, JS-based.
The one thing Jekyll has going for it is that it was extremely stable. We rarely have had to update stuff when new versions came out over the span of over 10 years.
I would like to retain that.
What are good website tools that we can easily adopt coming from Jekyll that don’t create major maintenance burdens down the line?
I don't have any personal experience with Astro itself, but from what I know it seems like it would be a good fit for replacing a Jekyll site if you're trying to stay within the JS ecosystem.
I've used astro a fair bit, and it's great. But I haven't used it enough to recommend it as long-term stable.
On the jekyll front, @janl have you tried:
npm run dev-site-with-docker
?
I also found Ruby a hassle, but this solved it for me.
Ref:
https://github.com/apache/pouchdb/blob/fe0f290ddc3660401b68b07b222a45d5800268e0/package.json#L26
npm run dev-site-with-docker
Oops, looks broken! Fix at https://github.com/apache/pouchdb/pull/9137
I've migrated a multilingual Jekyll site to 11ty and could be available for questions if you go that route.
Depending on desired planning, I could help you out building it in either Astro or Eleventy, both are great.
I'll throw Mastro's hat in the ring. Would be happy to help migrating the site.
If you can wait a little while longer, I would observe how https://zensical.org/ turns out.
It is developed by the same people as mkdocs-material, and that was excellent.
The first publicly released version was just two weeks ago but I have high hopes for it.
I had a quick look at eleventy, and it looks like it's almost a drop-in replacement: https://github.com/apache/pouchdb/pull/9146
Maybe worth pursuing?
Forgive me for being the devil's lawyer here, but the JS community is known for moving things very fast without much concern for greater stability or long-term maintenance. I understand the desire to use smt JS-based because, well, we're all JS devs, but "rarely have had to update stuff when new versions came out over the span of over 10 years" is smt very hard to have in the JS world (even Node.js have a LTS agenda that's way shorter than 10y).
If you're open for non-JS options, I'd recommend Hugo – IIRC, is the most popular SSG in the industry (if not the, certainly one of the top most), widely used, supported, maintained, etc., and it's written in Go, which has a pretty decent DX, and does not break things this fast.
PS: I speak as a person who loves/hates JS, been coding JS for 15+ years, and is very annoyed by these move-fast-break-things culture in the JS world, specially for projects that should live for 10+ years and does not change much very often.
@joeljuca Generally all true, but the JS community is not a monolithic. Look at 11ty's track record (e.g. this blog post). And what I can say in favour of Mastro is that it's only ~700 lines of code, there's only so much that's going to change, and you could also just copy those 700 lines to your repo to remove any dependency you don't want.