- From: Danny Ayers <danny666@virgilio.it>
- Date: Thu, 15 Jul 2004 00:34:26 +0200
- To: Laurian Gridinoc <laurian@gmail.com>
- Cc: www-rdf-interest@w3.org
Returning to your original example: ><rdf:Property rdf:about="http://NEW/#messageid" rdfs:label="messageId" >rdfs:comment="The unique Message-Id"> ><rdfs:subPropertyOf rdf:resource="http://purl.org/dc/elements/1.1/identifier"/> > ><rdfs:seeAlso rdf:resource="http://xmlns.filsa.org/xmlns.filsa.net/emir/0.2/#messageid"/> ><rdfs:seeAlso rdf:resource="http://www.openhealth.org/xmtp#Message-ID"/> ><rdfs:seeAlso rdf:resource="URN:ietf:params:rfc822:message-id"/> > ></rdf:Property> > > Would I be right in thinking that URN:ietf:params:rfc822:message-id etc. are other ways of describing a message id? To be used as: X hasMessageId Y If so, although owl:equivalentProperty might be an option, I'd be wary of using it because the alternatives might not be /quite/ the same. How about: <owl:ObjectProperty rdf:about="http://NEW/#messageid" rdfs:label="messageId" rdfs:comment="The unique Message-Id"> <rdfs:subPropertyOf rdf:resource="http://purl.org/dc/elements/1.1/identifier"/> <rdfs:subPropertyOf rdf:resource="http://xmlns.filsa.org/xmlns.filsa.net/emir/0.2/#messageid"/> <rdfs:subPropertyOf rdf:resource="http://www.openhealth.org/xmtp#Message-ID"/> <rdfs:subPropertyOf rdf:resource="URN:ietf:params:rfc822:message-id"/> </owl:ObjectProperty> In other words, http://NEW/#messageid shares characteristics of each of these, but perhaps not all, is a specialization. Assuming of course that you want to model the message ID as a property (of something) rather than a class in its own right, in which case do a quick search and replace on the above ;-) Again, if I'm reading the intention correctly, this is all working within the property(/class) hierarchy, and should stay comfortably within OWL DL, although I believe you will have to explictly type the referenced properties, i.e. something like: <owl:ObjectProperty rdf:about="http://xmlns.filsa.org/xmlns.filsa.net/emir/0.2/#messageid" /> I'm sure it could be done neatly with nodeIDs, but whatever, you get to stay in the TBox ;-) Cheers, Danny. -- Raw http://dannyayers.com
Received on Wednesday, 14 July 2004 18:38:44 UTC