Re: deprecated tags in Wilbur & Cougar -Reply -Reply

Mike Wexler (mwexler@mv.us.adobe.com)
Tue, 06 Aug 1996 16:15:09 PDT


Message-Id: <199608062315.QAA26338@orion>
To: www-style@w3.org, www-html@w3.org
Subject: Re: deprecated tags in Wilbur & Cougar -Reply -Reply 
In-reply-to: Your message of "Tue, 06 Aug 1996 15:58:40 CDT."
             <199608062058.PAA07537@predator.urbana.mcd.mot.com> 
Date: Tue, 06 Aug 1996 16:15:09 PDT
From: Mike Wexler <mwexler@mv.us.adobe.com>

I don't think that classes should be standardized by W3C or by UA authors.

I think authors and organization that create lots of documents should be
able to create there own classes.

I am currently working on a tool that exports FrameMaker documents to HTML.
It takes the styles specified in the paragraph and character catalogs
and generates a style sheet. It then will split a FrameMaker document (or
book) into many HTML pages that are linked together. All of the link documents
share the same style sheet. Documents that come from a common template will
also be able to share style sheets. 

It would be unfortunate if the W3C or UA authors decided to limit the
flexibility of CSS by specifying certain standardized classes. It would
mean at a minimum, that tools like mine would have to check for all the 
prespecified classes and avoid them (no telling if the author of a FrameMaker
document really meant paragraph tag FOO to mean the same as W3C's class FOO.

> If classes are going to be useful (by which I mean if they are going to
> leaed to reusable stylesheets), they need to be standardized.  The scope
> of the standard determines the useful scope of reusability of the
> stylesheet (a continuum from totally non-reusable-even-by-the-author to
> reusable by any author).  I think we will get de facto standard
> stylesheets courtesy of the major UA vendors in pretty short order.
> 
> Further, if classes are to be used to convey semantic information that
> is useful to tools (such as indexers), they must be standardized or the
> information the classes encode is unusable.

Classes should not be used to convey semantic information to indexers. This
should be left to element tags and other attriubtes.

> My own take on this is that we need a much larger standard set of
> elements and we need an effective social mechanism for getting new
> elements standardized quickly.

If we need a much larger set of standard elements, than lets create them.
Maybe the solutions is to get the UAs to start handling SGML documents with
DTDs specified by a URL. Than any organization could create a DTD that matched
the needs of its members. Some DTDs might be shared by a large number of
organiations. Others might be specific to a certain author.