W3C home > Mailing lists > Public > public-sws-ig@w3.org > July 2006

Re: mapping complex type to OWL concept

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Thu, 20 Jul 2006 13:20:19 +0100
Message-Id: <1982F409-2F88-4CCF-A32F-0FEC490B8A05@isr.umd.edu>
Cc: public-sws-ig@w3.org
To: "dhavalkumar thakker" <dhavalkumar@xsmail.com>

On Jul 19, 2006, at 1:31 PM, dhavalkumar thakker wrote:

> Hi,
> Thanks for the suggestion.
> I am sorry if I misunderstood you, but are you saying that if I will
> have my complexType such as :
> <complexType name="ArrayOf_xsd_string">
> <complexContent>
>   <restriction base="soapenc:Array">
>        <attribute ref="soapenc:arrayType"
>        wsdl:arrayType="xsd:string[]"/>
>   </restriction>
> </complexContent>
> </complexType>
> then I will always have trouble in successfully executing the OWL-S
> service?

Well, that's not quite what I said.

There are two ways to deal with an XML type in OWL-S...pass it on  
through or map it to an OWL or RDF description. Passing it on through  
has the advantage of, well, ease and fidelity. Mapping to OWL or RDF  
allows you to treat the XML type as a mere serialization format that  
represents something for which you have a better articulated theory  
for (encoding, for example, in OWL). So, if you *do* think of the  
type as a mere concrete representation for something you have a  
better theory for, then mapping makes sense. If you think of the XML  
type as being itself the direct representation of your object, then  
mapping to OWL doesn't really make any sense.

However, there is no current system that can handle complex types in  
OWL or RDF, which limits what you can do.

I suppose as long as you *can* get a URI for the type, there's  
nothing stopping you from popping it into OWL. I guess we don't have  
to assimilate all complex types to XMLLiteral, it's just that you'll  
have to do some escaping to represent the instance in RDF/XML and I  
don't know of any system that will do anything sensible with that type.

> because I am still not clear whether this  possible and if yes
> how?

Well, you'll have trouble because none of the toolkits that I know of  
are set up to handle this well.

> currently I am just specifying as a string(which doesnot make sesnse,
> however doesnt give any exception while executing which is  
> improvement)
> <profile:hasOutput>
>       <process:Output rdf:ID="fetchDataReturn">
>         <rdfs:label>fetchDataReturn</rdfs:label>
>         <process:parameterType
>         rdf:datatype="http://www.w3.org/2001/ 
> XMLSchema#anyURI">http://www.w3.org/2001/XMLSchema#string</ 
> process:parameterType>
>       </process:Output>
>     </profile:hasOutput>
> and the service executes successfully, but returns nothing(
> theoratically looks correct as the mapping is not proper)...

You might try just putting another URI in there that purportedly  
represents the Complex Type.

As for it returning but returning nothing...well...it's hard to tell  
what the problem is. If you invoke the service normally, what happens?

Received on Thursday, 20 July 2006 12:24:27 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:54:17 UTC