[dsdl-discuss] Re: Agenda for September

My understanding is that Part 6 is intended to fill in DSDL's gap with
ID/IDREF/KEY/
KEYREF.

This adds or uses declarative semantics that something in some context
is unique, or
that it is suitable for being used as a key. ID/IDREF is an edge case,
because really
to be useful it is information that belongs in the infoset (and
therefore implies
some kind of PSVI) but we have adopted "validation not augmentation" model
for the formalities of DSDL.

A) If James can go to W3C and define a subset of SGML and then XPath, I
don't know
why when he comes to ISO he cannot define a subset of XPath.

More seriously, if the purpose is for part 6 is not necessarily to
define a subset of XPath.
We could just reference XML Schemas, or even adopt STX (which I think
would be
good).

B) I agree that the first item looks too much like Schematron and should
be deleted.
For the second part, on identity and integrity constraints, these are
not rules, in the
sense that Schematron is rules. So it does not duplicate Part 3 and
should be kept.
Schematron can express the constraints, but not the semantic (unless we
reserve some
roles, which I am not keen on).

C) I agree that the issue of how to treat a document as if it belongs to
some other namespace is not an issue belonging in part 6, but in part 10
if anywhere. I think this is an important issue to consider: not so much
for making non-namespaced documents namespaced, but to allow versions
of documents.

D) "DSDL does not include a rational for using RELAX NG rather than W3 XML
Schema. What is the reason and why is it not stated in DSDL?"

They are working on XML Schemas 1.1 now, and there may well be some
future realignment
with XQuery. References we make now may be outdated.

In particular, there should be nothing in DSDL which sets it up as a
reaction
against XML Schemas, however much it may be that for some people.

Cheers
Rick Jelliffe

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)