The conclusion that style sheets are required for HTML/Web growth
does seem to be reaching consenus. The conclusion that HTML2.0
or HTML3.0 (as proposed) is an adequate means to represent web
information has not.
The people I talked with in Chicago, mostly SGML types, expressed
deep concern that either HTML was going to:
- get mutated to the point that it can not serve any needs well,
- or that several existing, real, and current information
encoding needs would not be met.
Either way, the web would continue to fragment and lose its
I believe that expressions of this concern are C.M. Sperberg-McQueen's
presentation, "HTML to the Max: SGML Intelligence to the World Wide Web"
and Jon Bosak's proposal. (available from
One problem of the current approach is how are you going to present
to author or reader something meaningful to help them adjust the
For example, our large authoring DTDs have over eight separate things
that end up as italic. It would do no one any good to simply say
"make all italics 14pt bold".
Whereas, it would help them to say
"make program variables bold".