Re: Generic Markup [was:Re: deprecated tags in Wilbur & Cougar]

Gavin Nicol (gtn@ebt.com)
Sat, 10 Aug 1996 13:37:50 GMT


From: Gavin Nicol <gtn@ebt.com>
Date: Sat, 10 Aug 1996 13:37:50 GMT
Message-Id: <199608101337.NAA24568@wiley.EBT.COM>
To: papresco@calum.csclub.uwaterloo.ca
CC: kmc@harlequin.com, marc@ckm.ucsf.edu, www-style@w3.org, www-html@w3.org
In-reply-to: <199608101324.JAA18696@calum.csclub.uwaterloo.ca> (message from Paul Prescod on Sat, 10 Aug 1996 09:24:23 -0400)
Subject: Re: Generic Markup [was:Re: deprecated tags in Wilbur & Cougar]

>I think it would be dangerous to build an information system that was
>neither based on a particular set of element names NOR had a mechanism for
>mapping to a standardized set of element names. That sounds like a recipe
>for chaos to me.

People have been doing this for some time now with less than chaotic
results. 

The thing you folks seem to be missing is that all SGML really does is
gives you a way of structuring your information. At one level, the
information structures for many (most?) things can be represented using
exactly the same information structure (for example, VRML, RTF). This
information is a tree (or a list of lists) of attributed nodes. For
such an information structure, SGML just gives you a (not particularly
pleasant) syntax.

The core to an application is *interpretation* of the tree. That is
why I think CSS is limiting: it has only a limited capability for
interpretation of arbitrary trees. 

If yu have a sufficiently powerful language for specifying the
symantics to be associated with the tree, you don't get chaos, you get
magic.