- From: Jeen Broekstra <jeen.broekstra@aidministrator.nl>
- Date: Wed, 14 May 2003 16:44:39 +0200
- To: "Roger L. Costello" <costello@mitre.org>
- CC: www-rdf-interest@w3.org
Roger L. Costello wrote:
> Oscar Corcho sent me an interesting idea - use owl:Nothing to represent
> EMPTY, e.g.,
>
> <rdf:Property rdf:ID="secret">
> <rdfs:domain rdf:resource="#Document"/>
> <rdfs:range rdf:resource="&owl;Nothing"/>
> </rdf:Property>
>
> A very interesting idea! Are there any drawbacks to this? Does it
> achieve the desired result of requiring secret to have an EMPTY range,
> e.g.,
>
> <Document>
> <secret/>
> </Document
This would not have the desired effect, though the thought is good.
Your ontology now indeed does specify that there can be no legal
values of a secret property. Sofar, so good.
However, it is an intrinsic part of the RDF data model that all
statements have subject, predicate *and* object. You simply cannot
have a property without a specified value. Thus, what your ontology
definition leads to is that you can not use the 'secret' property at
all, on any Document.
Having looked at your initial post, I would suggest a simple
modeling along these lines:
<rdfs:Class rdf:ID="DocumentTag"/>
<DocumentTag rdf:ID="secret"/>
<rdf:Property rdf:ID="tag">
<rdfs:domain rdf:resource="#Document"/>
<rdfs:range rdf:resource="#DocumentTag"/>
</rdf:Property>
<Document>
<tag rdf:resource="#secret"/>
</Document>
(Maybe I'm being far too practical here, but I'm just sharing what
came to mind :)).
Jeen
--
jeen.broekstra@aidministrator.nl
aidministrator nederland bv - http://www.aidministrator.nl/
julianaplein 14b, 3817 cs amersfoort, the netherlands
tel. +31-(0)33-4659987, fax. +31-(0)33-4659987
Received on Wednesday, 14 May 2003 10:47:07 UTC