Lists Home |
Date Index |
Tim's technologies-that-matter seems rather focussed on mass market
acceptence. SGML has very high and continuing penetration of
the industries it was designed for (large sets of technical and reference
documentation, batch-processed) and very little in industries it was not designed
for (casual documents).
That SGML does not "matter" in Tim's criteria is similar to saying
that the commands to CISCO routers do not matter, because
there is no mass awareness and other people use different systems.
When you have the need for the commands, they matter :-)
So if you are flying a plane, do the flight or operations manuals
matter? If the aircraft manufacturers and airlines have adopted SGML
(in some part because it was an ISO standard), does it matter?
SGML keeps chugging away for the same kinds of uses it was
good for 10 years ago: XML has allowed some nicer processing
systems for marked-up documents, but has little to offer
the people who had already successfully implemented SGML systems.
Perhaps we can say "What matters to Tim Bray" is de-emphasizing
SGML and emphasizing XML!
It would be interesting to know what XML would have brought
to Tim's OED project that SGML didn't have: is it just the ability to
not have DTD validation (or would just having content models
of ANY more often have done the trick?) How would XML
Schemas, RELAX or Schematron or AF have altered things
if they had been available then?
But in any case, we don't need to accept Tim's branding exercise.
XML 1.0 is SGML: ISO 8879:1996 specifically uses XML
as an example. ISO standards work best when based on
established technologies or industry profiles: so it is not
a bad thing that XML development is undertaken by
someone other than ISO. So if SGML does not matter,
it just means that ISO standardization is not important
in most cases, which is I hope is very true.
From: "Christopher R. Maden" <firstname.lastname@example.org>
> Of course, in many senses, XML itself is "just" a "convenient transport
> representation," with data coming from databases, generated from the state
> of some process, or otherwise not existing natively as pointy brackets.
I think we will see more of this kind of statement: soon the XML infoset will be
deemed to exist independently of the XML document. Rather than
markup languages being a discipline which challenges existing technology to include
rich narrative content, local encodings and local datatype notations, and arbitrary
annotations represented in plain text, XML will be dumbed down to be merely a
serialization format for databases (i.e. the DBMS of the major corporate participants
on the W3C, hungry to stick extra layers on their products to get sales).
When what matters is defined as what has mass acceptence, we don't get
> Of course, I think you knew this - but thanks for the leading question. (-:
Of course, I don't agree at all :-)
(writing as individual, and definitely not mattering)