WS-friendly XML profile?

The TAG is discussing a matter that came up at the Trout Pond session at our
last F2F, i.e., defining an XML subset/profile that hits some sort of 80/20
point for most users.  We talked for some time at the F2F about how the
whole XML corpus is simply getting too complex for ordinary mortals to
understand. I think the TAG discussion was triggered by the XMLP decision to
continue to forbid DTDs and PIs in SOAP messages.  See
http://lists.w3.org/Archives/Public/www-tag/2003Jan/0022.html
[search for 'xmlProfiles-29']

One sticking point seems to be whether the "profile" would include a way of
defining the ID attribute(s) without a DTD or schema.  Since SOAP uses ID
attributes, forbids DTDs, and doesn't assume that a schema has been
processed, this seems like it could be an issue on which we as a group might
want to take a position.

Dave Orchard, do you think a statement of the WS-arch position would be
well-received by the TAG? If not, let's not bother!

Does anyone care about this issue enough to discuss it here?  Does anyone
care enough to draft an "official" position of the WSA WG to discuss,
refine, and send to the TAG?

What are some other limitations of XML 1.x relevant to the WS world that
could conceivably be addressed in a future rev of XML?  I recall a couple
(not sure if it was from the F2F or the XMLP list] such as a) it would be
nice if different parts of an XML document (e.g., the SOAP headers and he
body) could use different character encodings; and b) it would be nice if
there were some way to parse a file (e.g. a logfile) full of well-formed XML
fragments without enclosing the whole mess in a top-level tag.  [probably a
non-starter for the XML geeks on the TAG :-).]  Thoughts?  Others?

Received on Tuesday, 7 January 2003 00:14:08 UTC