One thing I feel that the SGML proponents are missing are the following
three nice features of MIME richtext:
1. It's simple to parse.
2. Users can learn to use it quickly and hand-code in it.
3. It can be fairly robust in the face of hand-coding errors.
If a SGML DTD can be defined with these three features, great, let's see it.
But, if special tools are required to parse and especially to compose the
SGML documents then I don't think it would get as wide an audience (at
least in the short term), as much as I'd like to see something "clean". I
have no problems with using SGML (beyond my lack of knowledge of it :-).
The situation with hand-coding becomes even more important with international
character sets. I've been experimenting in recent months with Japanese and
Korean extensions to Metamail's richtext program, and quite a few hand-coding
and general encoding considerations have come up, especially where "<" is used
within multi-byte characters, and how character sets are switched.
A user doing hand-coding in a multi-byte character set may not be aware of the
fact that "<" occurs in their characters because they never see the "<" on the
screen themselves, and so won't manually replace it with "<lt>". Any format
that is chosen for formatted text in the future will have to cope with errors
in hand-coding, because hand-coding is the way a format will spread quickest.
I'd like to see examples of SGML that satisfy the above 3 properties if
anyone (Erik?) has any, together with a DTD-free description of the parser
(i.e. how to parse it if you don't want to create a full DTD-driven SGML
parser).
Cheers,
Rhys.