It may be that eyes glaze over when the letters SGML are uttered,
but we need it as a way of making rules and drawing limits. With
SGML I can parse my HTML docs and be confident that any browser
using the HTML DTD (and, let us hope, soon the HTML+ DTD) will
be able to handle my docs and not choke. They may even
be presented in a way that makes their structure clear by
All DTDs compromise between semantic and visual markup. The
exchange about white space touched on this point.
So the thing to do is enrich the DTD so that reasonable presentation
demands can be met, relying on semantic tagging rather than visual
tagging as much as feasible, then hide the whole SGML layer from writers
who don't want to hear about SGML. To the extent the markup of our docs
is semantic rather than visual, we're ahead.
After all, the non-Postscript alternative is RTF (snore) . . .
-- Terry Allen (email@example.com) Editor, Digital Media Group O'Reilly & Associates, Inc. Sebastopol, Calif., 95472