Lists Home |
Date Index |
Aww David, you just had to generalize and now we
have to tussle. :-)
1. SGML did not fail. In its day and in its
medium, it did quite well and lots of folks
made money including you. Its implicit processing
model and syntax options were not going to be accepted
on a stateless network and given the ubiquity
of Unicode. Otherwise, XML is just the bits
of SGML needed for that and the parts we all
knew worked without a lot of grief or deep
markup expertise. SGML was designed for a
different time in computer technology development,
when memory was expensive, CPUs were weak,
9-track tapes were the exchange medium,
and the delimiters reserved by each language
varied like women's shoes. In practice,
most of the excess features were never used
but they had to be there, or the standard
The tools were hard to write. Absolutely.
2. I never had a lot of trouble finding
markup errors in SGML, but maybe I didn't
use the right combination of all of those
usually unnecessary features. What are
you referring to?
XML succeeded based on the experience of
the parent language, the ubiquity of HTML
plus the perceived and real limits to applying
it, and somewhat, the web mania of the time.
If we tried the same thing today, we'd never get
it to a recommendation.
From: David Megginson [mailto:email@example.com]
Finally, I don't really see the need -- somebody suggests this kind of
thing every few months, and then it just dies quietly. It's also
worth noting that SGML allowed extensive syntactic abbreviation, and
SGML failed; XML forbade it, and XML succeeded. That's not the only
reason that SGML failed, of course, but it was a contributing factor
(SGML tools were just too hard to write, and markup errors were often
too hard to locate and fix).