Docs: Migrate from core.js to other framework

[zmzlois]

Not sure if my stance is simplistic but I'd love to hear creator's opinion.

As Denis Pushkarev written on Valentine's day, 2023, in the article So, what's next?, it was clear that he couldn't find sponsorship for core.js and the project is no longer maintained. A lot of big companies indeed still uses it, but at the same time, forward thinkers are diving in to try out new OSS frameworks such as Solid Start or Astro.build. I know new frameworks are not necessary to be used, but if this is what young people are attracted to -- why not?

I have asked several open source maintainers and contributors why they choose to contribute to open source and a lot of the time the answer is "Because this is cool and exciting."

I propose several new framework ideas to rewrite our documentation in hoping enthusiasts find our project exciting (it is both exciting and innovative) to encourage further collaboration within the OSS community and even attract people who are not from OSS community but excited to try out new technology. Many people's OSS journey starts with fixing a minor typo. And if we want to attract frontend developers to use backend technologies, we might as well see what could be the best out there?

If my idea gets approved, I will start rewriting our documentation from April onwards. Here are some thoughts on each framework:

1, Svelte
Writing less code but using languages people already known: HTML, CSS and Javascript. It align with Yiming's philosophy on "Don't learn the latest framework, learn the fundamentals" (I guess?) Because it's truly reactive and simple, if we ever decide to put blog examples into documentation (say something extremely interactive but built on Next.js), it could be both possible while retaining performant.

2, Astro
The reason for Astro is simple: stripping away Docusaurus. I know it was hairy to set it up...

3. Solid
   Several packages integrate Markdown, including one supports mdx compiler integrated with Vite.

• solid-marked
• vite-plugin-solid-markdown

While both Marko and Fresh are great, there are two major reasons why I don't think they are a good option:
1, strange..syntax.
2, Preact is ...relatively old? And Fresh is built on top of Preact, although it is performant but for documentation, until we have an extremely interactive component want to embed in the website, we don't need to worry about any performance-related issue?

[ymc9]

 Hey @zmzlois , thanks for proposing a rewrite. I agree that the current site is not stylishly designed 😄.

Just FYI, it's currently implemented with Docusaurus. It's not a new tool, but I feel it is quite solid for serving documentation: easy to configure, has a good variety of plugins, and generates an excellent experience score (from an SEO perspective). So far pretty happy with the choice. Btw, I guess core-js was noticed probably because it's a dependency of Docusaurus.

I feel Svelte/SvelteKit is worth trying for implementing a feature-rich web app. Haven't used Astro yet, but I'm interested in looking into it. For documentation sites, I think overall matureness and SEO friendliness are crucial (that's why I switched from docsify to docusaurus ...).

For the time being, I think reimplementing the doc site is not a high priority. Instead, maybe improving the quality of the content itself (copywriting & structure) brings more value to the developers. What do you think?

[zmzlois]

Copywriting & structure-wise... I do have some thoughts :)

FYI on how Astro looks like: every file feels like writing a markdown file + html, javascript is inside the front matter, and properties of markdown files are inside the front matter too. Without any extra configuration to implement markdown.


With a site as an example Eye of the storm editing time is about 6-8 hours for the first page