I read through the linked Grand Unification Proposal out of curiosity. I’m sure its author means well, but it shows a great deal of ignorance of all parts of the stack (HTML, CSS, JavaScript, accessibility, computer science, &c.) and why they are the way they are, and proposes extremely impractical solutions to the perceived problems, while introducing problems ten times as bad as the ones it solved.
It proposes ditching HTML and CSS in favour of JavaScript, but never seems to acknowledge that it also replaces JavaScript with a radically different beast that’s extremely incompatible with JavaScript, integrating an unspecified form of reactivity to the language, and shifting to some unspecified kind of lazy execution semantics. (Greenspun’s tenth rule of programming also applies to this mutation of JavaScript.)
It drops the web’s layout system completely, including all text handling of things like wrapping or inline blocks, in favour of something absolutely crippled that accidentally embeds the halting problem and unbounded mutual recursion. Dropping text layout would certainly lead to people reimplementing much of the existing system in user space, and that would be a multi-megabyte library that now everyone would be shipping. (This is not the only case where shifting it into JavaScript is a much bigger deal than people tend to imagine; CSS transitions are surprisingly fiddly to implement correctly in pure JavaScript; in the best case, your single line becomes a dozen obscure lines and some kind of extra state storage.)
Its examples of this brave new world are more verbose than the comparable HTML, elide critical details like z-ordering (which looks like it’s been done completely back to front, which would never fly for rendering performance), and get some basics of layout wrong, so that you’ll get collisions on tiny screens where the HTML a sane person would write using grid or flex (and perhaps even some float if they wanted!) would not have fallen over, and would be more maintainable.
There’s also a substantial dose of cognitive dissonance: “I cannot stress enough that it is imperative to adopt preexisting and proven concepts rather than reinventing the wheel” and bemoaning the inefficiency of the documents that people write because they have ten times as many elements as it seems they should need, while proposing (a) shifting such fundamentals as layout and even text layout into user space, which is plain insanity; (b) reimplementing hyperlinks in JavaScript, in the `onclick = "open('…')"` style; (c) replacing functioning accessibility and metadata systems with… some vague idea that’s going to be better, or something, but doesn’t actually say anything. (Least of all about the accessibility tree—I get the impression the author may not even have heard of that.)
No, I’m afraid this is all up a technically terrible proposal, with no real redeeming qualities at this time: most of the things it says are problems I disagree with, and most of its solutions are actually much worse problems. But I do think the proposer might like to learn some Lisp—you’re never going to convince people to replace the whole web stack like this, but understanding Lisp’s unification of data and code would definitely help develop some of these ideas into more practical things. Certainly any real unification of HTML, CSS and JavaScript would work best with S-expressions rather than some JSON- or JavaScript-like syntax.
Looks like the author suffers from Dunning–Kruger in a big way. Surely, he may know a lot, he has a MSc., not a stupid person at all, but it's obviously not enough.
It proposes ditching HTML and CSS in favour of JavaScript, but never seems to acknowledge that it also replaces JavaScript with a radically different beast that’s extremely incompatible with JavaScript, integrating an unspecified form of reactivity to the language, and shifting to some unspecified kind of lazy execution semantics. (Greenspun’s tenth rule of programming also applies to this mutation of JavaScript.)
It drops the web’s layout system completely, including all text handling of things like wrapping or inline blocks, in favour of something absolutely crippled that accidentally embeds the halting problem and unbounded mutual recursion. Dropping text layout would certainly lead to people reimplementing much of the existing system in user space, and that would be a multi-megabyte library that now everyone would be shipping. (This is not the only case where shifting it into JavaScript is a much bigger deal than people tend to imagine; CSS transitions are surprisingly fiddly to implement correctly in pure JavaScript; in the best case, your single line becomes a dozen obscure lines and some kind of extra state storage.)
Its examples of this brave new world are more verbose than the comparable HTML, elide critical details like z-ordering (which looks like it’s been done completely back to front, which would never fly for rendering performance), and get some basics of layout wrong, so that you’ll get collisions on tiny screens where the HTML a sane person would write using grid or flex (and perhaps even some float if they wanted!) would not have fallen over, and would be more maintainable.
There’s also a substantial dose of cognitive dissonance: “I cannot stress enough that it is imperative to adopt preexisting and proven concepts rather than reinventing the wheel” and bemoaning the inefficiency of the documents that people write because they have ten times as many elements as it seems they should need, while proposing (a) shifting such fundamentals as layout and even text layout into user space, which is plain insanity; (b) reimplementing hyperlinks in JavaScript, in the `onclick = "open('…')"` style; (c) replacing functioning accessibility and metadata systems with… some vague idea that’s going to be better, or something, but doesn’t actually say anything. (Least of all about the accessibility tree—I get the impression the author may not even have heard of that.)
No, I’m afraid this is all up a technically terrible proposal, with no real redeeming qualities at this time: most of the things it says are problems I disagree with, and most of its solutions are actually much worse problems. But I do think the proposer might like to learn some Lisp—you’re never going to convince people to replace the whole web stack like this, but understanding Lisp’s unification of data and code would definitely help develop some of these ideas into more practical things. Certainly any real unification of HTML, CSS and JavaScript would work best with S-expressions rather than some JSON- or JavaScript-like syntax.