RE: Two logical WSDL documents describing the same service

My last point wasn't very clear and TBH isn't well thought out. 

I was latching onto the <association> idea and musing that this could form the basis of a very useful service description: that an interface actually implemented or extended one or many other interfaces. 

Paul 



-----Original Message-----
From: Amelia A Lewis [mailto:alewis@tibco.com]
Sent: 12 January 2004 15:38
To: Downey,PS,Paul,XSJ67A C
Cc: dbooth@w3.org; www-ws-desc@w3.org; jacek.kopecky@systinet.com
Subject: Re: Two logical WSDL documents describing the same service


On Mon, 12 Jan 2004 09:56:43 +0000
paul.downey@bt.com wrote:
> AIUI you're saying that a service namespace must be unique and
> identify a single interface ? 

*I'm* not saying it.  I've been dissing it ever since the WG adopted it.

> If so, this sounds fairly Draconian and will make versioning an
> interface difficult:  We really need to be able to /enhance/ an
> existing interface without changing the namespace, i.e. make
> backwards-compatible changes without impacting existing users of a
> service. 

Too bad for you, under the current spec.  You can create a service with
another name (in the same namespace), but you can't implement a second
interface in the same service.

> I do like the sound of assertions to describe which interfaces a WSDL

??

> implements. Maybe it could be possible for WSDL to define a list of
> versions (Name/URIs) against an interface, or even against an
> individual operation ? Maybe this is something best implemented as a
> series of RDF statements ?

Completely lost me, sorry.

Amy!
-- 
Amelia A. Lewis
Architect, TIBCO/Extensibility, Inc.
alewis@tibco.com

Received on Monday, 12 January 2004 11:08:22 UTC