> However, I think (picking up on Nat's comments in http://www.vuw.ac.nz/\
> non-local/gnat/converters.html) that the HTML+ DTD should stand on its own
> feet as a usable DTD with "broad appeal". A common DTD encoding semantic
> information is obviously important for document exchange and reuse, and so
> we should strive to achieve this however difficult the task may seem.
Aaah, now this is a good point and the first time I've heard it
explicitly raised. If HTML+ is to be a DTD that encodes semantic
information for document interchange, then it should be exposed to
criticism and comments from a large group of people --- namely those
who are producing SGML documents at the moment. If we do less than
this, we run the risk of producing something sub-optimal.
> A DTD describing an "abstract presentation information" (which
> Nathan suggests) is no more or less than _another_ presentation
> format. Surely we can aim higher that that.
Surely we can, but do we really want to? The power of SGML isn't, as
is often touted, the ``removing of presentation from the markup'', but
really the separation of presentation and semantics. Obviously this
requires a presentation-free DTD, but doesn't preclude encoding the
presentation (what SGMLers call the ``style-sheet'') in another DTD.
What I envision is people writing their documents in whatever DTD they
feel best suits their semantics. Then they convert their SGML to
HTPL, a common presentation format for the web, and this is what is
put on the web. If we create a DTD that encodes portable semantics,
then by golly, they can convert their SGML to PSL (portable semantics
language) and make *that* available as well, but there's no good
reason to mix the two --- they serve different purposes. The
presentation version is for human reading, the portable version for
everything else.
Nat