Each approach has clear advantages:
Structured text -- can be easily displayed on a variety of display devices
with varying size, resolution, etc. Potentially very compact. Major
disadvantage is that visual design is heavily compromised. Appropriate
when content is more important than appearance; when bandwidth is
expensive; when many display devices must view a common document.
Page description -- presentation is preserved, visual impact and design are
communicated. Major disadvantage are the overhead of carrying all of the
visual design information; inability to adapt presentation to the output
We're starting to see major customers adopting *both*. Sun, for example,
is setting SGML and Acrobat as its standard documentation formats.
I could go on about this for a while, but I'd rather suggest that if the
designers of HTML abandon principles of structured text, they'll ruin it by
creating a standard that has the worst of both worlds.
I'll also point out that we're serving Acrobat documents over the Web
*now*, so this isn't a pipe dream. Check out <URL:http://www.verity.com/>.
I think designers who want a high level of control should stick with
Acrobat, Common Ground and their ilk, rather than putting pressure on the
HTML designers to break its paradigm. By way of disclosing a conflict of
interest, I don't want to leave out the fact that our engine is built into
Acrobat and is going into Common Ground.