Re: customer feedback, please (trackback)

Dan Connolly wrote:
> Can you clarify "quite-possibly invalid HTML"?
> Maybe this check-box applies?
>
> [ ] We rely on HTML that isn't XML, so neither of these
>         proposals works for us

It applies so thoroughly that I despair for my use case. Since it was easier
than a real sample, I checked the Technorati Top 100 blogs for ones
supporting Trackback. Of the 13 I found, 8 claim to be XHTML, four claim to
be HTML, and one lacks even a doctype. All 13 are so completely invalid and
ill-formed that a parser would stop and catch fire within a few hundred
bytes. I think if anyone decides they want the RDF out of Trackback data,
they are better off writing a special-purpose crawler that will regex it out
of the hot comments, parse it, and make it available to the rest of the RDF
world (though, along with being sloppy, my people are notorious for having
hissy fits when people use their metadata the way it's intended to be used,
so whoever does it better wear asbestos).

Phil Ringnalda

Received on Thursday, 18 March 2004 22:40:09 UTC