Like many Web-development bloggers, I've written about the importance of writing valid and semantically-rich HTML -- code that's heavy on self-description.
It's generally frustrating to evangelize semantic markup, though, because its advantages aren't immediately apparent. Why should people take time to put
<q> tags around quotes if Web browsers don't really do anything useful with them? The main point of semantic markup is to make documents easily understood by computers -- but computers don't seem to be doing anything exciting with the markup yet, on a large scale.
Well, now there's this thing, the W3C's Semantic data extractor. Pop in a URL, and it'll attempt to glean as much semantic information from the page as it can. It's a great way of visualizing what sort of data a computer can extract from a Web page.
What an outstanding idea. Developers should know about this.