Ah the sweet butthurt cries of a big iron dinosaur.
That was a joke. I was on the microcomputer side of that divide, where we ridiculed posixisms and told the "Code by the K-Loc" suits to go plow themselves.
The thing people miss about HTML is right there in your post, the idea that it's presentation in the document... and you're right, they did start piling hack upon hack with HTML 3.2, but that wasn't HTML 2/earlier's fault, it was the fault of people thinking that HTML was to say what things look like.
That was NEVER supposed to be its point from day one. And the reason it told "what came before" to sod off is a hefty chunk of this.
You say it with the intrinsic markup part, which is inherently presentational and NOT what HTML was originally for.
HTML was for saying what things ARE, structurally and grammatically so that the user-agent could best convey that meaning. It's why the original browser TBL made actually had stylesheets for different device targets. The entire problem he was trying to fix was the plethora of documents PRETENDING to be hypertext that weren't conveyable easily across different devices, because they was no list of tags based on actual writing norms.
Didn't matter if it was teletype, graphical print, tty, 80x25 MDA text, the 21x22 text of the VIC=20 Linus Torvalds cut his teeth on, or the 1152x864 display of the NeXT workstation TBL was sitting at. The user-agent could use the tags based on professional writing norms to convey that meaning.
That's what separates HTML from SGML, XML, RDF, OWL and so forth. If HTML didn't provide tags with meanings that UA's would understand predefined -- aka "semantics" by the dictionary meaning of the word (something RDF lacks) -- there would have been no reason to use HTML instead of just SGML, XML, RDF, or any of the other "container" formats that actually LACKED literary semantics.
A good comparison is that HTML is a data format in a SGML container, just like how MKV, AVI, MOV, MP4 and so forth are containers for codecs like H.262, huffyYUV, etc, etc. Or how WMF is a container format for different windows media data like BMP, DIB, ICO, CUR, etc.
That's where those research bodies you mentioned "career educated" themselves into obscurity and back-room server geek wankery. Tossing around words like intrinsic, extensic, ontology, not because it makes anything easier or better, but just to implement an obtuse legalese-like verbiage designed to maintain an L33t classism. Aka, "screw the normals".
Which is why what came before HTML was effectively stillborn, and what came after (both in and outside HTML) has often been equally irrelevant in the mainstream.
It may have been arrogant, but it sure as shine-ola wasn't done out of ignorance. Rather than spending 40 years circle-jerking each-other with big words and stuff utterly irrelevant outside of the classroom, TBL looked at a problem and implemented a solution.
And it's even there in your opening paragraph. The ALLEGEDLY semantic junk that are RDF and OWL focus on the machine, but not on delivering content to users or actually providing meaning. As such, they are NOT semantic no matter how many decades of going NOWHERE they have bandied the word about.
That's the biggest flaw in those technologies is that they focus more on the "machine" than they do what such formats should exist for -- delivering content to users in as useful and accessible a manner as possible!
But then I know, crazy me expecting words in the esoteric theoretical sciences to actually be used for their dictionary meanings. Don't expect that in the web programming world; closure, semantic, empty, void...
As I've said a few places it's like the bullshit of the media saying they can't call Brock Turner a rapist because he wasn't convicted of rape. No, he was convicted of sexual assault with a foreign object... you know, the dictionary definition of rape.
Anyhow, used properly -- aka ignoring the idiocy that was HTML 3.2 / 4 Tranny -- and going back to HTML's original purpose practicing separation of presentation from content, you actually get the best of both worlds in terms of intrinsic vs. extrinsic... because for the most part you should NOT be chaning the HTML to change the presentation.
In fact, if you choose ANY of your HTML tags -- well, other than DIV/SPAN as hooks -- for the purpose of presentation, you are choosing all the wrong tags for all the wrong reasons.
Presentation is NOT supposed to be ANY of HTML's business, and that goes back to when it was created. It was only the garbage browser makers made that got shoved into the spec with 3.2 any-old-way that pissed on that original intent. An intent 4 Strict with CSS tried to drag us kicking and screaming back to.
Thus your entire "achilles heel" paragraph misses the entire reason HTML was even created!
"All that work" was ignored for good reason, it was theoretical bullshit that wasn't actually going anywhere. Nothing it was trying to do actually solved problems that needed solutions, and simply seemed to exist for a certain community to sit there holding a giant verbal circle-jerk.
There's a reason HTML drives the web, and RDF and its kine are rarely seen outside of proprietary file formats and back-room server-nerds who need to get a bit more sunlight. In no way do they provide a format that a single UA standard could be built to convey meaning and accessibility regardless of device capability or user limitations.