Re: [xml-dev] Semantics and the Web: An Awkward History

["Simon St.Laurent" <simonstl@simonstl.com>]

On 9/14/2021 3:49 AM, Marcus Reichardt wrote:
> Thanks Simon for sharing this fascinating read.
You're welcome!

> Not quite getting the blues near the end where you seem to decry the loss of relevance of markup technology. Wouldn't it be more productive and less depressing to assess where we stand and formulate future goals?
You absolutely can look on the bright side, but the relentless "always 
look on the bright side of life" is usually an excellent way not to 
notice things that have broken or are in danger of breaking.   This 
perspective was also shared by at least one of the reviewers of the paper.

From my perspective, we don't especially have future goals.  I'm 
delighted that JSON, YAML, and who knows, maybe KDL (pronounced 'cuddle' 
- https://kdl.dev/ ) have redirected the endless demands that XML is 
only useful if it looks like OOP objects.  I think, though, that we 
squandered opportunities that could have made for better use of markup, 
and see little energy remaining even for maintenance, much less grand 
new adventures.

> IMO, XML is just fine as it is, it's not talked about much because it does its job, and I'm not seeing anything that could replace it any time soon, or even at all.
That resembles Tim Bray's long ago "XML should be like ASCII, too 
ordinary a thing to have conferences" perspective.
> That's because the world where standards were defined based on broad consensus, academic canon, and customers demanding standards in an ecosystem of multiple commercial players has largely ceased to exist.
I agree.  I'm not sure how strongly it ever existed, but it existed 
enough to be convincing for a while.

> But the original goals of markup, as set forth in eg the foreword to The SGML Handbook, haven't been achieved when seen in a broad perspective. As in, will future generations, or even our future selves be able to read content created in the last decade, as it was presented to the original audience? Or will any discourse that may lead to future events, such as climate crisis, media concentration and polarization, pandemics, etc. be left for future researchers to decipher in ways much worse than is possible now with medieval or Greek/Roman antique sources? A humanist may chime in here to tell us about a historic scholar who took it upon him/herself to define a lingua franca or general techniques for preserving text during the renaissance or antique epochs; that's how important our work is IMO.
>
> We don't even have to appeal to digital humanism to see value in markup; any person starting a work of text large and small is confronted with a choice of just using glass pearls/blue pills offered in the cloud, or go all the way using sustainable bona fide text formats such as SGML- or TeX-based ones. For those who know, and have experienced the document format monopoly of the 1990th, this choice is also about self-respect, and putting our money where our mouth is.
There was a moment, roughly 1998-2003, when I thought we were going to 
achieve those goals across a large swathe of digital content, as "XML by 
default" seemed to be getting there.


> In this context, I'll admit I have only read your text transcript rather than the YouTube video. Have we even begun to explore declarative means for contemporary UI and media presentation btw?
XUL?  XAML?  SMIL?  They mostly linger as zombies, with some 
implementations still running but development wandered elsewhere. They 
all fit into the arc of this piece neatly.


> I'm aware your article is about semantics in markup, but I'm suggesting to drop the semantic high horse stance, and try to reassess what the purpose of a text format is. I've worked in high profile RDF/SPARQL projects, but the very idea of separating structure and presentation (in eg poetry or any other artistic text, or educational text for learning a language), let alone use microsyntax to enshrine one particular way of separation is outright laughable. I'm appreciating postings by R. Costello here, acting as kind of an impressario to keep the list alive, but I feel the topics addressed have little or nothing to do with digital text. Rather, they seem to discuss all things coming out of the mouth of W3C. Speaking of which, isn't HTML's stagnation, and CSS's out-of-control complexity to compensate for it, a direct consequence of W3C's fixation on XML?
I tend to see - and attempted to make clear in the presentation - the 
stagnation of HTML and the abandonment of XML as having far more to do 
with programmer preferences for imperative models than W3C fixations on 
markup.  The sequence where the specs ballooned was driven by the W3C's 
bizarre Semantic Web priorities and members' efforts to shove in as much 
sugar as as possible for programmers.

I don't see CSS as especially complex.  Again, not imperative, so not 
amenable to programmers. Its structure came from well before the W3C 
thought of XML, and it more or less survived an effort by the XML 
community to replace it with XSL(T/FO).

I don't share your opinion that separating structure and presentation is 
laughable.  I do it daily?

Thanks,
Simon