This has been pointed to in a number of places, but the Google Code examination of real world web markup use is fascinating (if geeky) reading. Certainly an insight into what once can do with large sets of raw data when you begin to slice it (the slicing is neither good nor evil, though the commentary might be - though the delicious relaxed language is a nice touch).

Like many amateurs, I'm blushing to realize that if they had happened to hit some of my pages (particularly those I've left behind from earlier lives, but probably current ones also) they'd find a number of common (read 'newbie' errors).

More validation would help my pages, but at least I don't leave MSWord output tags all over.... Textmate does just fine for most work, thanks.

It's interesting to see how the various <meta> elements and <a> extensions have panned out (though there are no actual percentages shown throughout). nofollow seems to have a bit of traction by comparison to the others, but tag seems less used than I would have initially expected, though after reflection on its rather specific purpose at the moment, not surprisingly so.

Note: the graphs require SVG support (e.g. Firefox). First time I've knowingly seen SVG used so nicely.

Posted by esinclai at January 25, 2006 09:28 PM |