W3C home > Mailing lists > Public > public-ws-semann@w3.org > February 2007

RE: What's an Attribute Declaration component?

From: Jacek Kopecky <jacek.kopecky@deri.org>
Date: Tue, 13 Feb 2007 17:07:33 +0100
To: Jonathan Marsh <jonathan@wso2.com>
Cc: SAWSDL public list <public-ws-semann@w3.org>
Message-Id: <1171382853.8536.82.camel@localhost>

Jonathan, 

many thanks for doing this, I suspect we may count it towards our
implementations. And I really enjoyed your poetic blog entry. 8-)

Some more remarks below.

On Fri, 2007-02-09 at 18:47 -0800, Jonathan Marsh wrote:
> 2)      I visually inspected wsdl-xslt results, found them to be to my
> understanding of the spec, and added them as Baselines against which
> other implementations can compare.  Double checking my understanding
> of what˙s correct might be wise ;-).

I'll try to take a look, or get someone from the WG to take a look. 8-)

> 3)      The Attribute Declaration item below prevents me from
> implementing support for test case 8.  I don˙t see a straightforward
> solution wrt where the annotation will appear in the WSDL 2.0
> component model.  At the least it seems to require adding components
> to the model rather than simply annotating existing components.

This is now our CR issue 5 [1]. I expect that instead of adding
components to WSDL, we will talk about XML Schema components. So this
test case will not manifest in the WSDL component model. Anyway, we'll
let you know when we decide.

[1] http://www.w3.org/2002/ws/sawsdl/issues/CR-20070126#x5

> 4)      The modelReference, liftingSchemaMapping, and
> loweringSchemaMapping attributes are defined as sets of URIs.  By this
> I infer that the order of uris listed within them is not significant.
> Wsdl-xslt inserts them into the component model in their order of
> appearance.  The canonicalization algorithm does not yet sort them, so
> an implementation that didn˙t keep their order would result in false
> negatives in the interchange comparison algorithm.  This probably
> isn˙t worth fixing until the problem actually manifests.

I expect you are right that it's not yet worth fixing. I don't expect
that we would test for equivalence of WSDL files annotated with same
annotations in different order, it wouldn't seem very useful.

Best regards, and thanks again,
Jacek
Received on Tuesday, 13 February 2007 16:07:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:58:46 UTC