<p>A page that is
</a>, all done in CSS!
So, yanno. Welcome to our dystopian internet hellhole, and all that…
I was in middle school when I started to make websites, which meant my coming of age was defined by the idea that I could make websites and put them on the web for anyone to see. It was intoxicating in part because it was an open platform without gatekeepers deciding what methods of building websites were worthy. If you knew some HTML and could upload it to a web server, you were in.
Jeremy Wagner on the web.
I’m reminded of every time I’m salty about, e.g., React on Mastodon and someone has to slide into my mentions to defend the framework. The defenses are always developer-centric, and even more than that, always established-career-developer-working-in-an-opsdev-startup-envrionment-centric. They’re never things like “easy for anyone to pick up and implement!” or “fast-loading, semantic, and accessible across devices” which, y’know. Given I learnt to code in the girlweb heydey of “it’s XHTML compliant or you’re CANCELLED” sure is… different.
Also see: every shitty single-use webapp project on Github that requires its own fucking VM and/or Docker image preloaded with esoteric libraries to run. Seriously, go fuck yourselves. I know y’all don’t think PHP is ~cool or whatever, but at least if I download something written in it I can be 99% sure it’s going to run reasonably politely on my existing webserver and not force me into shelling out for an extra VPS…
On the birth and death of the semantic web.
I confess I… was kinda into the semantic web. Back In The Day, when I used a homebrew system rather than WordPress, my entire blog was even rendered in XML, styled for display to humans via XSLT transformation.1 This was back in the XHTML heyday. I always hated XHTML—it’s basically the worst of both HTML and XML and I’m glad it “lost” the standards wars—but I liked the clean separation of data and presentation as represented by the XML/XSLT combo.2
It had, of course, some obvious downsides, the main one being that it would crap itself and fail if it encountered even slightly malformed data (as opposed to HTML, which is endlessly fault tolerant). It was also very easy to machine-parse, which is both good in the sense of interoperability but, conversely, also bad in the sense of interoperability, depending on your stance on content silos, web scraping, and plagiarism.
Still. It was neat, and clean, and I do kinda miss it. Modern “alternatives” like microformats and JSON whatevers and and such and so on always seem like… pale and compromised imitations.
- A technique I stole off the then World of Warcraft website, which implemented this, as I discovered when I once tried to view-source in order to steal graphic elements for my guild site. [↩]
- I even implemented it in a commercial project that required different format outputs for data. It was a pain to make the underlying framework actually do what I wanted, but once I’d done that, it was way easier to just write new transform stylesheets for, say, Word-versus-Excel-versus-browser output. [↩]