Re: Semantics of WSDL vs. semantics of service

On Mar 15, 2006, at 11:36 AM, Shi, Xuan wrote:

> Jacek,
>
> Thanks for your explanation. If you agree that semantic annotations 
> have no
> direct relation with WSDL elements,

Whatever this is, I don't agree. I imagine there will be semantic 
annotations that have direct relations with what certain WSDL elements 
describe (however briefly).

> then why don't you create a separated
> and independent document

Do that if you like. Or put it in the WSDL. Why not?

> to describe the *meaning* of your services?

Well I don't want (necessarily) describe THE *meaning*, I want to 
describe my service in a variety of convenient ways. I think it will 
often be more convenient to have a richer WSDL document.

Note that OWL-S does the "annotative" approach (I.e., owl-s documents 
point to wsdl documents with no modifictation of the wsdl). Which is 
fine. Not the only way to go, but fine.

However, it does NOT buy you any magical separation that resolves your 
"problem". These are notational, not substantive, differences.

> That's
> just my suggestion which was proposed as the so-called OSRR approach 
> for
> SWS. I also demonstrated already that such approach can be deployed by
> either SOAP or REST services without question.

Yeah, but in that debate you totally failed either to understand or 
respond to the fact that WSDL (e.g., the wsdl describing the sparql 
protocol) can do the latter "without question". So, you have no 
advantages whatsoever (technically).

> Service semantics have no relation with WSDL-

Pif and fle.

> which can be deprecated as I
> demonstrated

You have asserted this. Ad nauseum. But sheer assertion does not 
establish your case.

[snipped more of the same]

Even more of the same won't help, btw. Actually responding to, or even 
acknowledging, the rather devastating critiques would at least allow 
moving the conversation forward.

Cheers,
Bijan.

Received on Wednesday, 15 March 2006 22:14:13 UTC