Stop Standardizing HTML
pfignaux writes with an interesting view on the place of centralized standardization in modern browsers. From the article: "When HTML first appeared, it offered a coherent if limited vocabulary for sharing content on the newly created World Wide Web. Today, after HTML has handed off most of its actual work to other specifications, it's time to stop worrying about this central core and let developers choose their own markup vocabularies and processing."
Instead, the author proposes that CSS, Javascript+DOM, the W3C's accessibility framework, and Web Components are sufficient to implement the rendering of smaller, domain-specific markups.
You ask for a companion programming language and at the same time propose eliminating Javascript. I see a contradiction in there.
The author's proposal sounds suspiciously like he has fallen for the seductive path of elegant generalizations(that are too theoretical for the ugly details to yet be visible) instead of confronting the ugly details that our current attempt at standardization has made visible...
To be sure, the sausage-by-comittee that tends to result when you try to standardize something is quite ugly and takes ages to settle down; but if the proposal is "Just let people use whatever shims they want for everything" you haven't really solved the problem, just comitted yourself to standardizing a suitably powerful interface for the shims to sit on, along with giant piles of shim-dependent code that crawlers and any other applications that break the shims' assumptions won't be able to make the slightest sense of.
Heck, for maximum elegance in the core standards, we could just replace virtually everything with the "Object" tag, and let people embed whatever they want, or abandon this 'HTML' nonsense entirely and just make Native Client the standard, freeing developers to implement pretty much any conceivable structure, from a legacy browser engine, to a Flash client, to a TECO interpreter built entirely out of Minecraft redstone logic, as the shim for their 'site'. A glorious age of unfettered freedom!
HTML5 is the response by a bunch of whiners that normal xhtml is "too hard." Yes it's too hard to remember to close your tags. It's too hard to remember to put quotes around attributes. Why are humans checking your syntax? Have the danged computer check your syntax.
"Pave the cowpaths" is an excuse to appease a bunch of zealots that are hellbent on pushing their personal preferences and egos into a standard rather than designing something that is quick/easy to parse and universally render across platforms. It's only going to get worse as the standard is never completed over the next decade.
XML serialization of HTML sucks. It's verbose, and it's ugly. But it's effective because it's well defined and it leaves very little room for interpretation.
Honestly, I'd like to see two standards. One, is XHTML5 Strict that follows the XML serialization. This will be left to the big boys who have real work to do. The other standard would be an extension to MarkDown to allow CSS customization with classes and ids. This would allow the path the cowpaths crowd to get things down as fast as possible and keep the verbosity of XML out of their way.