- From: Sandy Gao <sandygao@ca.ibm.com>
- Date: Thu, 8 Feb 2007 12:38:09 -0500
- To: Erik Wilde <dret@berkeley.edu>
- Cc: www-xml-schema-comments@w3.org
- Message-ID: <OF7B379899.0D435F04-ON8525727C.0060203B-8525727C.0060E0FA@ca.ibm.com>
Erik,
My quick reading of the request suggests that it's similar to one of the
issues the WG considered earlier. See [1]. The WG considered the request
and decided that the spec is correct and the clauses do not need to
change. The WG also observed that this issue is a source of confusion and
wants to clarify it in a note. See [2].
Can you confirm that this is the same issue as the one you are raising?
Note that there are 2 steps involved in Identity Constraint validation.
1. Form the "key-sequence". In this step, it is indeed allowed for a
"field" to match nothing. (Hence the key-sequence is incomplete.)
2. Verify the key-sequence. In this step, for "unique", it's OK to have
incomplete key-sequence, but it's not OK for "key".
So there isn't a single answer to "are incomplete key-sequences allowed".
It depends on the variety of the identity constraint.
[1] http://www.w3.org/Bugs/Public/show_bug.cgi?id=2543
[2] http://www.w3.org/Bugs/Public/show_bug.cgi?id=4202
Thanks,
Sandy Gao
XML Parser Development, IBM Canada
(1-905) 413-3255
sandygao@ca.ibm.com
www-xml-schema-comments-request@w3.org wrote on 2007-01-18 10:24:49 AM:
>
>
>
>
> hello.
>
> i am proposing a clarification of the section about how identity
> constraints are validated. when i look at section 3 of the validation
> section http://www.w3.org/TR/xmlschema11-1/#d0e20523, i see the
> following text:
>
> "For each node in the ·target node set· all of the {fields}, with
> that node as the context node, evaluate to either an empty node-set
> or a node-set with exactly one member, which has a simple type."
>
> i would say this clearly allows having fields selecting optional
> items. maybe this could be made more explicit, however, because
> implementations differ. eclipse accepts xs:field selecting an empty
> node set, whereas other implementations complain about that (they
> seem to assume that the selector of a field must always select a node).
>
> maybe a note could make the spec easier to understand here.
>
> cheers and kind regards.
>
> erik wilde tel:+1-510-6432253 - fax:+1-510-6425814
> dret@berkeley.edu - http://dret.net/netdret
> UC Berkeley - School of Information (iSchool)
>
>
Received on Thursday, 8 February 2007 17:38:23 UTC