Re: rdf:property expressions in DAML

On Thursday, June 19, 2003, at 04:05 AM, Monika Solanki wrote:

> A Trivial question.

Not so trivial, actually.

> I have noticed in Process.owl and also CongoProcess.owl that certain  
> properties have been expressed as rdf:property rather than  
> owl:DatatypeProperty or owl:ObjectProperty, though the range is not  
> owl#Thing For e.g.

That range shouldn't matter. rdf:Property is the super property of both.

> rdf:Property rdf:ID="locateBookOutput">
>   <rdfs:subPropertyOf  
> rdf:resource="http://www.daml.org/services/owl-s/0.9/ 
> Process.owl#output" />
>   <rdfs:domain rdf:resource="#LocateBook" />
>   <rdfs:range rdf:resource="#LocateBookOutput" />
>   </rdf:Property>
>
> and
>
> - <rdf:Property rdf:ID="nameBookInCatalogue">
>   <owl:domain rdf:resource="#BookInCatalogue" />
>   <owl:range rdf:resource="http://www.w3.org/2001/XMLSchema#string" />
>   </rdf:Property>
>
> Could there be a specific reason behind doing this or does it not  
> matter?

The specific reason is that our coverter did that, And the specific  
reason it did that is that daml:Property is used in the daml files.

The biggest thing that this does is it makes all those ontologies  
definitely in OWL Full (syntactically, at least). In general, the DL  
and Lite species of OWL require you to be very explicit of the type of  
just about everything and to be mutually exclusive with your set of  
Object and Datatype properties. OWL Full doesn't either restriction,  
and many DAML+OIL systems, though they had the latter restrictions (for  
certain semantics) would let you infer from range or from use which  
type of property you had.

We didn't do anything more than an exact convertion of the files this  
round, partly due time and partly due to the fact that both daml-s and  
OWL are somewhat moving targets. It's on my todo list to figure out  
what species of owl DAML-S and useful DAML-S decriptions can reasonably  
fit it, and whether that buys you anything.

Cheers,
Bijan Parsia.