Re: SGML on the web (was: when will CSS rule)

| Wrong again. HTML *is* SGML. Therefore it is obviously possible to
make
| SGML-based languages that are "all style" or "part style" or "all
content",
| depending on your preference. Considering the gross, complex HTML
code
| required to do some things, I think it makes a lot of sense to make
| format-oriented SGML-based languages for format-oriented sites.

Nope HTML != SGML by any comparison function you can throw at it.  HTML
is SGML + mild style + etc...

| This is *not* an accurate paraphrase of my paragraph above which does
not
| mention validation at all. Let me say it again. It is useful to use
SGML in
| situations where you *never* intend to validate *anything.* Usually,
| however, once people move to SGML they figure they might as well take
| advantage of all of its features, instead of half of them. That will
| probably change if it becomes the lingua franca of the Web.

Then there is no need to use SGML, any thing else in the world will
work, without SGML though, there is no standard to force people to live
by, even if it does allow them to express their data in any way, it
still requires that they use the rules to write their DTD.  HTML is no
different, its an application of SGML, it means it can be described
using a DTD, but it also applied several other rules that become a
standard that binds authors and browsers, search engines and the
content in a simpler form than SGML permits.

Received on Wednesday, 20 November 1996 22:59:44 UTC