W3C home > Mailing lists > Public > xmlschema-dev@w3.org > August 2005

Re: implementing redefinitions

From: <noah_mendelsohn@us.ibm.com>
Date: Fri, 19 Aug 2005 10:00:13 -0400
To: Kasimier Buchcik <kbuchcik@4commerce.de>
Cc: "Henry S. Thompson" <ht@inf.ed.ac.uk>, Michael Kay <mike@saxonica.com>, XML-SCHEMA <xmlschema-dev@w3.org>
Message-ID: <OF4BC1C63D.F0EEE75C-ON85257062.004C283C-85257062.004CED0A@lotus.com>

Kasimier Buchcik writes:

>  Is there an effort scheduled to clarify redefinitions for XML Schema 
1.0 as well?

Interesting question.  I think it's fair to say that most of the WG's 
efforts have been targeted at Schema 1.1, but that said, the guideline has 
been to not change the function but only the explanation with respect to 
Schema 1.0.  That goal is complicated by at least two factors (1)  because 
the Schema 1.0 Rec is at best vague and some would argue outright 
contradictory or inconsistent on some points, there is disagreement about 
what represents a change and what doesn't...almost anything we write in 
1.1 will be viewed as a change by someone, which is one of the reasons 
we're aiming it for 1.1 (2) as one who has attempted to draft frameworks 
for clarifying composition, I can tell you that it is tempting to take 
some approaches on edge cases that are provably different than any 
reasonable interpretation of Schema 1.0.  I don't think these would change 
the results for most combinations of schema documents in practice, but 
they clearly represent changes.  The WG is still debating what approach to 
take, and truth be told, much of the work has been on hold recently while 
we are trying to get higher priority matters resolved. 

When you get into composition, you find out that it also is closely 
related to the mappings for individual components, I.e. the specifications 
of how particular properties in a component are set from a schema 
document.  Consider, for example, complex cases in which there is 
inheritance of a simple type facet from a redefined type.  The facet is 
described with the component, but the redefine mechanism is general to all 
components.  The point is that to get this straight you have to touch, or 
at least carefully review, large swaths of the existing text in the Rec. 
That's a big job.  We're hoping to do it, but right now the likely result 
is a Schema 1.1 Rec that clearly explains either one of the plausible 
interpretations of Schema 1.0, or something very close.  I thin that can 
be used as de-facto guidance for those who may wish to stick with Schema 
1.0, but it probably wouldn't be normative for them.

The above is just my opinion, not an official position of the workgroup.


Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142

Kasimier Buchcik <kbuchcik@4commerce.de>
Sent by: xmlschema-dev-request@w3.org
08/19/2005 05:16 AM
        To:     "Henry S. Thompson" <ht@inf.ed.ac.uk>
        cc:     XML-SCHEMA <xmlschema-dev@w3.org>, Michael Kay 
<mike@saxonica.com>, (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Re: implementing redefinitions


On Wed, 2005-08-17 at 16:00 +0100, Henry S. Thompson wrote:
> Hash: SHA1
> Thanks for this clear and succinct statement of the situation, which I
> think the XML Schema WG should take into consideration as it tries to
> clarify the whole schema composition story for XML Schema 1.1.


Is there an effort scheduled to clarify redefinitions for XML Schema
1.0 as well?

I'm currently unsure if the fact that nobody clarified the test
results yet (which you can find at the beginning of this thread),
means that the people who actually can comment this issue
are still on holiday or at meetings, or if this means that they will
not comment it, for whatever reasons. Please let me know if and how
this issue was noticed by the WG. I don't intend to apply any
pressure, but need to know if it is reasonable to suspend the
implementation of redefinitions on my side, in order to try to
find out what a schema processor _should_ do here.

What to do if the WG will not clarify the problems I see with
redefinitions? Are implementations then free to implement it
in a way they think is suitable? Does this mean that we
have another official area of implementation-dependant behaviour


Received on Friday, 19 August 2005 14:00:33 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:15:30 UTC