RE: service identification (was: Re: Web Services Definition and XML)

Sanjiva,

This exact question is one of the reasons why I'm continuing to hit my head
against the "what are we" wall.  I clearly understand how you and WSDL view
web service definitions.  "Service– a collection of related endpoints."

Is a web service a resource or is a web service a collection of resources?
I'm not sure that the WSA has really thought through this issue, and I don't
think the WSA is automatically required to accept the WSDL definition.  This
is the first time the community has had the opportunity to visit the web
service definition in a public forum.  The WSA should explicitly ratify a
definition.

The issue that I see is that the web consists of resources identified by
URIs.  WSDL defines web services as a resource identified by QName, and the
resources within the QName are identified and accessable by URIs.  So a Web
Service is NOT URI addressable as it currently stands in WSDL, rather a
(WSDL) Web Service endpoint is URI addressable.  Note the distinction
between whether the web service (the qname) versus the web service endpoints
being URI addressable.  Again, the Web Service in WSDL does NOT send/receive
XML, it's the web service endpoints.

Looking at this another way, a WSDL service is a qname, which is not
normally dereferencable.  And most namespace names that are dereferencable
return either html or a schema definition - certainly not the web service
itself.

I took a trundle through some various definitions.  Don Box [1] refined his
definition of a web service to typically involve http and xml.  I think
almost everybody thinks of web services in terms of xml, protocols, uris.
The WSDL definition has the ports being related to xml, protocols, uris; not
the service itself.

The subtle distinction between these definitions is whether a web service is
a resource, or a web service is a collection of information including
pointers to resources.   Perhaps a web service is both a "thing" and a
collection of "things".

I believe a key aspect of the web is uniformity of interface to resources,
particularly through URIs and XML.  I can live with the current definition
that mentions URIs and XML.  You and others may disagree, and
disagreement/discussion is good and natural.

Please note that I believe that providing a wide variety of information
about a web resource, particulary Web Services, is very useful and usually
required.  So I'm not arguing against any of the functionality of WSDL.

Cheers,
Dave

[1] http://msdn.microsoft.com/msdnmag/issues/01/11/WebServ/WebServ0111.asp



> -----Original Message-----
> From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]On
> Behalf Of Sanjiva Weerawarana
> Sent: Tuesday, March 05, 2002 11:19 AM
> To: www-ws-arch@w3.org
> Subject: service identification (was: Re: Web Services Definition and
> XML)
>
>
> I meant to send this to the list, not just to David.
>
> Sanjiva.
>
> ----- Original Message -----
> From: "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>
> To: "David Orchard" <david.orchard@bea.com>
> Sent: Tuesday, March 05, 2002 8:23 AM
> Subject: service identification (was: Re: Web Services
> Definition and XML)
>
>
> > Hi David,
> >
> > > Like I argued for URIs, I will also argue for XML in our
> definition.
> This
> > > is a show-stopper.
> >
> > I didn't follow that thread, but it seems to me a service should
> > be identified by a QName and not a URI. The reason being that a
> > service defines a bunch of things belonging to a single namespace
> > and being able to point to specific aspects of that service is
> > very useful. That's the approach that WSDL took.
> >
> > If this has been discussed already then I would appreciate a pointer
> > to the thread.
> >
> > Thanks,
> >
> > Sanjiva.
> >
> >
>
>

Received on Tuesday, 12 March 2002 12:49:36 UTC