W3C home > Mailing lists > Public > www-ws-desc@w3.org > May 2007

Re: [Bug 4380] Proposal for Assertion InterfaceMessageReference-0042

From: Amelia A Lewis <alewis@tibco.com>
Date: Mon, 21 May 2007 14:37:48 -0400
Message-ID: <758e9357cb1628f4850f7106f1c77778@xerom.na.tibco.com>
To: "Ramkumar Menon" <ramkumar.menon@gmail.com>
Cc: "Jonathan Marsh" <jonathan@wso2.com>, www-ws-desc@w3.org


On 2007-05-21 14:04:17 -0400 "Ramkumar Menon" 
<ramkumar.menon@gmail.com> wrote:
> The specification of a uniqueness constraint is incompatible with the
> optionality of the attribute.
> I am trying to understand the rationale of this explanation, but with 
> no
> luck. :-( Maybe I am missing something here.
> Would it be possible to elaborate more on this?

It's possible that we missed something, in discussing this.  We were 
under the impression that optional attributes could not be used in the 
selectors of unique constraints.

Further, the validation that can be done using the unique constraint, 
when the target attribute is optional with defaulting rules that are 
also inexpressible in WXS, is less than ideal:

<wsdl:operation> <!-- using default, so input-output -->
     <wsdl:output />
     <wsdl:input />
     <wsdl:output messageLabel="Out" />
     <wsdl:input messageLabel="In" />

This can't be caught with the uniqueness constraint (which is rather 
the purpose of the uniqueness constraint).  If we can't catch such 
cases, then adding the unique constraint to the XSD is apt to suggest 
that all that needs to be checked is the result of that, when in fact 
it is entirely inadequate as a consequence of the optionality of the 
attribute and the rules for defaulting its value when it is missing.

Note that the above fails for other reasons (not expressed in the 
schema, but expressible in the assertions of the component model).

Amelia A. Lewis
Senior Architect
TIBCO/Extensibility, Inc.
Received on Monday, 21 May 2007 18:38:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:55:04 UTC