RE: definition of web services in "web services architecture"

> -----Original Message-----
> From: Sanjiva Weerawarana [mailto:sanjiva@watson.ibm.com]
> Sent: Sunday, December 01, 2002 8:12 AM
> To: WS-Desc WG (Public)
> Subject: Fw: definition of web services in "web services architecture"
> 
> 
> 
> Hmm. Do we adopt the perspective that Mike is referring to below?
> If so it seems to me that we'd have to make the R120 proposal from
> Arthur into a normative one .. contrary to what we plan on doing
> now as I recall.
 
Hi Sanjiva,

I certainly wasn't trying to make work for the WS Description WG <grin>.

FWIW, I was referring more to R81 and the definition of an end point in the
Requirements draft:

"EndPoint (AKA Port)
[Definition: An association between a fully-specified InterfaceBindingand a
network address, specified by a URI [IETF RFC 2396], that may be used to
communicate with an instance of a Web Service. An EndPointindicates a
specific location for accessing a Web Service using a specific protocol and
data format.]"

Could you elaborate on why my explanation of what I thought was the
relationship between the WS-Arch draft document and WSDL 1.1 would require
changes to WSDL 1.2?  I personally am happy with R120 as drafted, and
recognize that this is a very difficult are to get consensus on in the W3C.

I believe the rough consensus in the WSA WG is:  make sure that "important"
web services resources such as endpoints are identified by URIs; this allows
conventional WSDL technologies to simply treat them as identifiers but
allows semantic web technologies to make assertions about them and make
inferences based on those assertions.  This helps move WSDL 1.1 technologies
closer to the conventions of the Web (clearly a "good thing" from the
perspective of many, especially the W3C leadership and the TAG) without
"breaking" anything too badly, and it provides a basis on which semantic web
technologies can build as they mature.

As the whole topic of URIs, URI references, XPointers, and the relationship
of all this to RDF has occupied much of the TAG's attention for the last
year and is still quit controversial, I don't think we in the web services
area want to attach ourselves too closely to any one perspective at this
point.  The best way forward, IMHO, is in keeping our stuff aligned with the
Web and the compatible with the semantic web while keeping options open and
requirements broad.

Not speaking in any official capacity for the WSA WG!  

Mike

> >
> > I can't speak for the working group's intention vis a vis WSDL 1.1.
> > Nevertheless, the general purpose of the definition is to 
> look forward to
> a
> > "reference architecture" incorporating WSDL 1.2 than to be backward
> > compatible with everything we now think of as a "web 
> service".  This does
> > not mean that WSDL 1.1 can't be used to describe "web 
> services", just that
> > it is best practice to use URIs whenever feasible to 
> identify web service
> > components so as to maximize compatibility and 
> interoperability with the
> > rest of the Web.  WSDL 1.2 adopts this perspective.
> >
> > In light of this, could you suggest some specific 
> clarifications in the
> > text?
>

Received on Sunday, 1 December 2002 09:01:06 UTC