W3C home > Mailing lists > Public > www-ws-desc@w3.org > May 2003

Re: using RDF (OWL) in WSDL

From: Jacek Kopecky <jacek@systinet.com>
Date: 28 May 2003 10:30:09 +0200
To: Arthur Ryman <ryman@ca.ibm.com>
Cc: WS-Description WG <www-ws-desc@w3.org>, www-ws-desc-request@w3.org
Message-Id: <1054110609.10840.27.camel@localhost>

Arthur, I hadn't understood the action the way you presented it below,
but I'll look into that too. Here's my initial thought:

When we present the RDF mapping of WSDL, the OWL description of that (in
other words the OWL schema for WSDL) will naturally include the
properties we cannot express in XML Schema.

The area where the interaction of RDF and WSDL needs some work is
embedding RDF in WSDL, which is how I understood the action, even when I
was writing it in IRC.

Your alternative, i.e. to allow rdf:Description or rdf:RDF in the XML
Schema schema for WSDL, would simplify the reuse of existing RDF
processors a bit, but then the syntax would usually look like this:

<wsdl:message name="Foo">
 <rdf:Description rdf:about="<targetNamespaceOfFoo>#message(Foo)">
   <ns:responsibleArchitect rdf:resource="mailto:jacek@systinet.com"/>

The rdf:Description element would not be tied to its parent message in
any way then and it could be simpler if we just allowed rdf:RDF as a
child of wsdl:definitions or introducing wsdl:semantics element which
would contain elements in any other namespace that would be assumed to
contain semantic descriptions of the stuff defined in the parent WSDL

With my proposal, wsdl:rdfDescription would be just another
extensibility element and the interaction with existing RDF processors
would involve the WSDL processor rewriting the wsdl:rdfDescription
element to the rdf:Description element (see above) which is a trivial
thing; and then passing the rewritten element to the RDF processor.

What do you think?

                   Jacek Kopecky

                   Senior Architect
                   Systinet Corporation

On Tue, 2003-05-27 at 18:46, Arthur Ryman wrote:
> Jacek,
> I thought the action was to look at using OWL in our WSDL
> Specification to encode the assertions that could not be expressed
> using XML Schema. It would be useful to have the additional assertions
> encoded in a formal way so we could avoid natural language
> ambiguities, and maybe even have an OWL processor validate WSDL
> documents.
> Your proposal for putting RDF in WSDL is interesting, but wouldn't it
> defeat existing RDF processors? Why not just change the WSDL schema to
> allow RDF elements?
> Arthur Ryman
> Jacek Kopecky
> <jacek@systinet.com>
> Sent by:
> www-ws-desc-request@w3.org
> 05/27/2003 11:59 AM
>         To:      
> WS-Description WG
> <www-ws-desc@w3.org>
>         cc:        
>         Subject:      
> using RDF (OWL) in WSDL
> Hi all,
> I was tasked to check how RDF statements (including OWL statements)
> could be represented in WSDL. Here's what I found:
> RDF defines two XML elements which can be used to contain RDF
> statements
> - rdf:RDF and rdf:Description. As far as I can see, no XML Schema for
> these elements (I've found only one at [1]) puts them into any of the
> substitution groups for WSDL extensibility elements; therefore these
> elements cannot be used as WSDL extensibility elements.
> As I think RDF statements should be embedded in WSDL using
> extensibility
> elements, I propose that we create an element wsdl:rdfDescription in
> the
> substitution group globalExt. The semantics of this element would be
> the
> same as of 
> <rdf:Description rdf:about="URIref to the current WSDL component">...
> Example:
> <wsdl:message name="Foo">
>  <wsdl:rdfDescription>
>    <ns:responsibleArchitect rdf:resource="mailto:jacek@systinet.com"/>
>  </wsdl:rdfDescription>
> </wsdl:message>
> In English, the architect responsible for the WSDL message Foo is me.
> What do you think?
>                   Jacek Kopecky
>                   Senior Architect
>                   Systinet Corporation
>                   http://www.systinet.com/
Received on Wednesday, 28 May 2003 04:30:29 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:06:30 UTC