Another way to think about this is, people were already thinking along these lines and tried to build something, and that thing is AMP. If you want to build something that avoids the failings you see in AMP, you're going to need to think hard about how your plan is different.
- that everyone has to serve all their content from a single domain. I only suggest serving optional Javascript from a single vetted repo.
- that the distribution repo should be owned and controlled by one if the worlds largest advertising companies.
- that the solution only should cover mobile.
For your other points:
* The AMP project is now part of the OpenJS foundation, along with Node, Electron, and others: https://openjsf.org/blog/2020/06/23/openjs-world-day-one-hig...
* AMP is not specific to mobile, though it did start that way; there are sites that serve all of their pages in AMP format to both mobile and desktop users.
(Disclosure: I work for Google, speaking only for myself)
Meaning that right now it isn't that way, right?
> The AMP project is now part of the OpenJS foundation, along with Node, Electron, and others: https://openjsf.org/blog/2020/06/23/openjs-world-day-one-hig....
For some reason everyone still seems to associate amp with Google, and the only times I can remember finding AMP pages are when I search with Google.
If a site is technically AMP but served from a non Google domain, I admit I haven't noticed it.
Correct, the current spec doesn't allow that. But, as I said, I think that is something that could reasonably easily change if many sites wanted to publish with vanilla CSS and HTML.
> For some reason everyone still seems to associate amp with Google, and the only times I can remember finding AMP pages are when I search with Google
Google started the project, and is still by far the biggest contributor. But Bing also supports AMP, caching and serving AMP pages: https://www.bing.com/webmaster/help/bing-amp-cache-bc1c884c
If so it has some huge downsides that overshadowes it.