This begs the question: Is it healthy that one of the most important technologies today (web) is also very complicated (=expensive) and thus effectively limiting number of players.
If we look at other technologies, the same thing happens there. How expensive would it be to write a new OS from scratch? A new C++ compiler? A new UI toolkit? These technologies are all incredibly complex.
Heck, consider how many projects are just frontends for LLVM. The LLVM IR claims to be well documented [0], but AFAIK it lacks a formal spec as well as multiple independent implementations.
But vendors getting together and saying "no, we're just not going to compete with native platforms in this space" seldom happens. As a proponent of the open Web, I think that's a good thing.
If we want to move forward as a society, we better structure our tech in the simplest possible way. The difference with cars, airplanes, etc is that in a capitalistic world, an overcomplicated browser spec actually helps big players to keep their monopoly. That's what's wrong here.
What im reading here is a preference for web technologies to be frozen as they were since 2000.
One first insight: we don't need the "semantic web" because Machine Learning can recover semantics. (This is also how Google's webcrawler views the web, so why not apply it to browsers too? People mix up semantics and formatting too often for the "semantic web" to be a useful concept).
Another thing to realize is that overcomplicated systems benefit from factorization, i.e. splitting functionalities into different modules with their own responsibilities. This doesn't just apply to code, but also specifications.
A better factorization will immediately solve the "monopoly" situation, because now many players can develop many modules, and browser vendors can simply combine module implementations at will.
What it will also solve is security issues. A modularized architecture is much easier to keep secure than a system where all functionality has been thrown together on one big heap without structure.
You can say that browsers are (if the developers are smart) already structured in a modularized way, but that is not the point, because the internal boundaries of the modules are not openly available in any kind of specification.
Personally, I long for a world where formatting a document or showing a video on a computer screen is not considered "high tech". And this might be the way towards it.
Note that an "extreme" interpretation of this would be to replace the browser by a simple virtual machine. In that case, every web-developer is their own browser vendor, and they can mix and match existing libraries at will. People will then probably object that the user loses power, and accessibility is lost. But we live in a world of Machine Learning now, so any "structureless" rendering can be easily restructured by an appropriate tool. E.g. an image containing text can be OCR'd and read aloud by a tool. This would also make these tools more robust against formatting hacks. And machine learning can do many more useful things, like removing ads.
Okay so here's the thing. Web pages do a lot more than just 'formatting a document' or showing a video. Any serious talk about the web platform has to acknowledge the that that it's an 'application platform' that is more than just an .rtf file.
So what I wanted to say is that the "high tech" is currently not as much in the functionality, as it is in the software engineering effort to reach that functionality.
Also, the paragraph after the sentence that you quoted addresses web-apps, if that's what you are concerned about.
It's hard to see exactly what you're asking for. Web specs are already quite modular. The modularity could be improved, but it's not like nobody's trying.
One way would be to create a browser architecture from truly independent pieces that would allow one to pick implementations from different sources. That would split complexity and hopefully increase choice.
To get this off the ground one would need an architecture and a functional reference implementation, which is a lot of work. If you get that, others will improve it from an OK state to greatness, but it must be end to end OK to start with for people to notice