- From: Henri Sivonen <hsivonen@iki.fi>
- Date: Sun, 6 May 2007 22:12:34 +0300
- To: Jukka K.Korpela <jkorpela@cs.tut.fi>
- Cc: www-html@w3.org, public-html@w3.org
On May 6, 2007, at 17:36, Jukka K. Korpela wrote: > On Sun, 6 May 2007, Henri Sivonen wrote: > >> In some cases, inferences may be useful if they are predominantly >> right and wrong only a relatively rarely. > > You haven't explained > a) why inferences should be made from existing usage at all In the case of class='copyright', I don't know. No one has presented consumer use cases to me. In the general case, if it is agreed that making inferences of new content is useful, making it from existing content as well helps bootstrap the new inference technology by letting it leverage the existing network of content. >> Considering that research shows that "copyright" is one of the >> most common class values, it is a reasonable hypothesis that it is >> used to mark up copyright notices and that there is demand for >> markup for copyright notices. > > No, it's just a guess, and a rather wrong guess. I thought it was a reasonable guess. If research shows it is a wrong guess, I won't argue otherwise. My point was mainly about pre-empting research by axiomatically and categorically forbidding speccing inference from existing content. > It's also based on a wrong assumption that class names should be > interpreted using an English dictionary. The point is exploring *if* they can be interpreted using existing practice as a guide. This is subtly different from just using a dictionary: if research showed that a non-word string was consistently used to denote something useful, a dictionary would not have to be involved. Since it is improbable that the string "copyright" would appear accidentally without someone thinking of the concept of copyright while writing the string, it is reasonable to assume that the string is motivated by the concept. -- Henri Sivonen hsivonen@iki.fi http://hsivonen.iki.fi/
Received on Sunday, 6 May 2007 19:12:51 UTC