Congrats! Considering HTML but mostly CSS complexity is expanding all the time as we speak (discussed on the HN front page right now [1]), are there points to be addressed at the likes of W3C, Inc and github.com/whatwg to make your work easier or even possible? Such as versioned CSS specs/CSS profiles (subsets), test suites aligned with CSS or HTML versions, test suites usable without JS (eg web-platform-test for CSS requires in-browser scripting last I checked), formal specifications/better algorithmical layout descriptions, etc? Are you targetting specific HTML versions such as WHATWG's preview drafts/snapshots?
Good luck with your project! Maybe naming your sponsors can help to get even more contributions and contributors?
I can offer some help/advice with HTML parsing, but I guess you have figured that one out and are following the procedural parsing description, whereas my expertise would be the SGMLisms on which it is based (and its minute re-construction from the HTML5 spec).
but mostly CSS complexity is expanding all the time as we speak
That's the biggest problem facing new browser implementations, and IMHO it's deliberate. Probably 99% of sites out there do not need any of the new crap that gets constantly added to the "standard", but continued propaganda by those constantly in power will practically guarantee that they will, under some guise of "progress" or "modernity". WHATWG is basically controlled by Big Tech (and was created because they didn't like the W3C's approach to stability) and popularised the idiotic oxymoron "living standard". Constant churn is how they're trying to maintain their monopoly, because they have the sheer manpower to outrun competitors.
I agree in principle; a reasonable spec that doesn't change all the time is what's needed for supporting next-gen browser development (as in inspire a new generation). But that's not what's happening at all; collaborators of github.com/whatwg even welcomed the rejection of HTML profiles/subsets and target device classes when the idea was brought up by MS (to save IE, but still ;)
But the ones responsible for the piece of shit that is CSS are actually W3C, Inc. Not that the bad looks of Google having them by the balls is better than your scenario eg the impression W3C on its last legs receiving money to keep the illusion of a multi-party effort of web standards.
A lot of CSS seems excessive but I, for one, am glad with a lot of the additions over the years. CSS tables are great (better than flexbox for many layouts IMO), CSS masks combined with gradients and other background tricks save us from manually positioned SVGs that have to be just the right size.
Recent additions include the :has pseudo class, which I've wanted for a while, a subgrid layout to alleviate the problems nesting grids can cause, scroll snapping should make those annoying Apple style scrolling web pages less of a problem to deal with, and there are tons of other great features out there as well.
With these, we can finally get rid of the janky mess that we needed 10 years ago. No more magical "width: auto auto" for centering text. No more lists of "display: float" to get dynamic sizing.
The newer the standards are, the better the documentation is in my experience. Older CSS had a lot of shitty implementation defined behavior that has since been documented to correspond to what most browsers seemed to do anyway. The newer standards seem to be written with an actual intend to be implemented consistently.
You could build almost any modern web page in Microsoft FrontPage twenty years ago with enough carefully crafted GIFs, but I think we all agree that CSS has improved for the better since then. If it weren't for all these "living standards", you would barely be able to use the web on your phone. I remember the PDA internet era, back before the iPhone made showing real websites on your smartphone a thing developers became aware of. It was not good.
Good luck with your project! Maybe naming your sponsors can help to get even more contributions and contributors?
I can offer some help/advice with HTML parsing, but I guess you have figured that one out and are following the procedural parsing description, whereas my expertise would be the SGMLisms on which it is based (and its minute re-construction from the HTML5 spec).
[1]: https://news.ycombinator.com/item?id=36375582