I’m currently looking into adorning this website with a bit of semantic
information. Because that’s where the web’s heading isn’t it?
Well, I’ve briefly looked into schema.org, which describes a reasonably simple set of types of information that can be directly embedded into the various HTML tags. So that’s a plus, but there’s something missing: There is neither a type for “source code” nor a type for “description of software”. There’s “blog” and “blog post” and all that - and the format is extensible. But still, that seems like quite a big omission to me. There’s “software application” and “review” and “article”, so you could make it an “article” about a “software application”. Well, yeah. That’s not very elegant, is it?
(As an aside I find it rather funny how the programmers designing and implementing those kind of things - other examples would be XML and most GUIs - completely forget about their own profession. I mean, obviously only catering to programmers would be bad too, but not at all thinking about programmers when programming something isn’t quite right either, is it?)
Also, if you really want to present all the information you have about whatever it is you’re publishing, the HTML tag soup gets rather - soupy (well, that didn’t work). Maybe it would be better to a have a scheme where you basically just provide the site template as an extra page - but instead of the usual content you just describe what kind of content there would be, if this was the actual page. Something like:
<h1>title</h1> <div id="foo"> blog-post-content <span id="bar">author</span> </div>
And the other side (search engines for example) would fetch this template and parse it and apply it to your page. That way, you wouldn’t have to rework all your markup if you want to include semantic information. Also, the question of what to specify would go away, because you would only say what it is, what you’re publishing - and not have to worry about whether to include the author on every post or only on the main “blog”-element, or only on the main “website”-element and so on.
Aside from schema.org, there are countless other ways of including semantic information into HTML pages. And I can’t read about them and compare them because I keep falling asleep. It’s so boring.
Sometimes I long for the days of the font tag. Things weren’t perfect, but nobody said they would be. Everything is so right in today’s web. You separate semantics from presentation and separate that from that and load everything asynchronously and everything defers everything somewhere else and “don’t use
<b>” and it’s all so static and it’s still a huge mess. But now with stricter rules.
So, fuck semantic tags.
Was that a post that started as a happy story about how I’d like to include semantic tags into my homepage and ended in a somewhat diffuse rant against - uh - progress?
You bet it was.
Previously: Tacit Programming in Racket
Next up: Electric Dead Trees
The machine thinks that the Web-Log entries <font color="green">, 26th Ludum Dare - Postmortem, and Purely Functional Games might be related to the topic so eloquently discussed above. The machine is sometimes right.