- From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
- Date: Tue, 6 Aug 1996 20:31:52 -0400
- To: marym@Finesse.COM (Mary Morris), www-style@w3.org, www-html@w3.org
At 05:01 PM 8/6/96 -0700, Mary Morris wrote: >I'll agree that authors should be able to create their own classes. In >fact, I second this opinion heartily. I don't disagree. I think that most of us just want to standardize _some_ classes within _some_ communities. That wouldn't prevent you from inventing your own. Some classes would only be used in a particular document, others across a web site, others across a corporation, others across an industry and some by all of us. >I for one would be willing to support a default definition for a >specific class keyword *if* it isn't overriden by the author or >explicitly rejected by the user. Class doesn't describe how things look, but what they _mean_ -- the semantics of the tags. Groups (i.e. corporations, industries, the web as a whole) will design class dictionaries that assign meanings to classes in the same way that regular dictionaries assign meanings to natural language words. If you choose to adhere to a particular dictionary, you will have to adhere to the semantics of the classes. That may or may not impact the default rendering of the class -- it's up to the UA designer. You have to leave it up to the UA designer and end user, because only they know the constraints and features of their display. So on a particular platform, a CLASS="phone" might be represented as a link which dials the phone. >I don't believe that stylesheets as we define them to control >presentation should be used to manage meta-information. I don't think that there is any relationship between style sheets and meta-information. I also don't think that CLASSes and META are very related, although they can both be used by indexers. >I'm also going to step in here and ask some general questions about >people's opinions on this matter. In theory, we define additional >invisible information or meta-information with the <META> tag and >sometimes with the <LINK> tag. However, this is at least one grass >roots movement that says that this is an awkward, cumbersome, and >in general lengthy way to define information. They go into an alternate >suggestion at http://www.identify.com/ I personally am open to suggestions of improvements to META, but IDML is a) not SGML compatible, b) not as extensible as META c) specific to a particular (though large) market segment (commerce). It might be worth inventing some dedicated tags that duplicate some of the things that META does, but they would have to be done in an SGML compatible manner so that they can be easily merged with HTML. Paul Prescod
Received on Tuesday, 6 August 1996 20:33:28 UTC