W3C home > Mailing lists > Public > public-swbp-wg@w3.org > April 2004

Re: [ALL] Human-friendly syntax for communicating OWL fragments

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Wed, 07 Apr 2004 15:38:31 +0100
Message-ID: <40741267.3010408@hplb.hpl.hp.com>
To: "Uschold, Michael F" <michael.f.uschold@boeing.com>
Cc: Sean Bechhofer <seanb@cs.man.ac.uk>, Frank van Harmelen <Frank.van.Harmelen@cs.vu.nl>, public-swbp-wg@w3.org

Uschold, Michael F wrote:

> 
> Jeremy: you said that there are problems with trying 
   to do this with OWL-Full. Can you give a simple example
   that shows the problem?
> 


Here's a very simple one:

Saying that Pat Hayes is the author of RDF Semantics may be said in RDF 
with a single triple:

<http://www.w3.org/TR/rdf-mt> dc:creator "Pat Hayes" .

To say that in OWL AS we may well end up with something like

individual(
    <http://www.w3.org/TR/rdf-mt>
    annotation( dc:creator "Pat Hayes" ) )

The difficulty is that the mapping rules automatically will generate
dc:creator rdf:type owl:AnnotationProperty .
as well as the triple we want and some other fairly harmless triples.

A different way of saying a similar thing in the AS would be:


individual(
    <http://www.w3.org/TR/rdf-mt>
    value( dc:creator "Pat Hayes" ) )

And here we would need to have dc:creator as a datavalued property, giving 
the triple:

dc:creator rdf:type owl:DatatypeProperty .

Under the OWL Full semantics these two are not semantically identical (in 
fact neither entails the other), and neither is entailed by what might 
actually be written in a concrete document which might just be the single 
triple.

Jeremy
Received on Wednesday, 7 April 2004 10:47:33 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:17:11 GMT