Re[2]: Clark's commentary

On Monday, January 07, 2002, 4:25:53 PM, Norm wrote:

NW> / Stephen van Egmond <svanegmond@tinyplanet.ca> was heard to say:
NW> | Clark proclaimed that we "should be free to stab the SGML
NW> | community, what's left of it, in the back.", claiming that SGML
NW> | compatibility is now of much lesser importance.   This seems like a
NW> | highly dubious claim - what constraints does the SGML influences
NW> | inflict upon us?

Lack of structured attributes springs to mind.

NW> In a sense, we've already stabbed them in the back with namespaces. As
NW> far as continuing influences, the first thing that comes to mind is
NW> determinism. After that, I think a solution to the general problem of
NW> character entities in a post-DTD world will probably require a
NW> complete break with SGML.

Or a web compatiiity annex two

NW> | Clark recommended that XML Namespaces and Infoset be merged into the
NW> | XML core spec, and DTDs be supplanted. Architecturally, is this an
NW> | improvement, or no?

NW> Probably. I think I'd (personally) be in favor of an XML 2.0 if *and
NW> only if* there was agreement beforehand that XML 2.0 would be XML 1.0
NW> + Namespaces + the Infoset + XML Base. (And not one iota more or less;
NW> no other changes. None. Not one.)

If you had added "a solution to the ID problem" in there I would have
been right behind that suggestion.

NW> Without the proviso that there would be no other changes, the XML 2.0
NW> effort would turn into a 90 person committee with everyone wanting to
NW> add or subtract their own favorite or most hated features and the
NW> effort would fail or produce something unusable.

Well, that is certainly a possible outcome but if its a likely
outcome, what does it tell us about the cohesion or otherwise of the
XML market?


-- 
 Chris                            mailto:chris@w3.org

Received on Monday, 7 January 2002 10:46:05 UTC