Re: deprecated tags in Wilbur & Cougar -Reply

Paul Prescod (papresco@calum.csclub.uwaterloo.ca)
Thu, 01 Aug 1996 17:27:14 -0400


Message-Id: <1.5.4.32.19960801212714.0075ed10@csclub.uwaterloo.ca>
Date: Thu, 01 Aug 1996 17:27:14 -0400
To: galactus@stack.urc.tue.nl (Arnoud "Galactus" Engelfriet), www-html@w3.org
From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
Subject: Re: deprecated tags in Wilbur & Cougar -Reply

At 08:28 PM 8/1/96 +0200, Arnoud "Galactus" Engelfriet wrote:
>In article <1.5.4.32.19960801134613.00748d04@csclub.uwaterloo.ca>,
>That would only partially solve the problem. By putting information
>about the *contents* of marked-up text in a style sheet, you're
>actually lessening the power HTML can offer.
>
>Style sheets are for layout.
>HTML elements are for contents.

CLASS is for content, just as elements are for content. Therefore the debate
about whether to make something an element or a class need not involve
stylesheets at all. The questions we should ask are: 

"Does this data class have content model implications?" If so, an element is
more appropriate. Classes cannot have specialized content models or occur in
different content models than their elements.

"Is this class specific to a small vertical market?" If so, a class is more
appropriate. Elements must be supported by all tools, everywhere (until all
tools support generic SGML).

"Is this class important enough to change the DTD and incur these costs?" 

"How soon is the new DTD going to be available?" Since classes can be
deployed more quickly, a class may be more appropriate.

"Is this concept experimental?" Classes can be dropped more easily than
elements, so a class may be more appropriate.

"Are we unsure?" Classes can be turned into elements more easily than vice
versa, so when in doubt, a class may be better.



Note that none of these questions involve style sheets. Style sheets are a
way of attaching visual representations to particular elements _or_ classes.

 Paul Prescod