Good point.
OK, here's the real reason I don't want to put ISO10646 in the SGML decl
for 2.0: I don't know how to do it, and I don't have tools to test it.
The SP parser by James Clark is almost certainly the answer to both
of these problems, but there's a third: I don't have time to mess with
it.
This is all clearly a cop-out, but here's a fourth reason: we don't
need it. As everyone has pointed out, ISO8859-1 and ISO10646 are
completely consistent, and there's no problem upgrading. Nothing specified
in the 2.0 spec (or used in most browsers) requires anything more than
ISO9959-1.
So the main reason I resist the change is that this is very late in
the release cycle for this sort of thing. There just isn't enough
"burn in time" on this idea. Heck: the dang thing is in last call, after
all!
For example: imagine the surprise of all the folks out there happily
using SGMLS on their HTML files to -- god bless them -- validate their
HTML. They download the final, released standard public text, and
SGMLS craps out cuz it doesn't grok the document character
set. Sadness.
ISO10646 all in good time.
Dan