rdfs:domain and rdfs:range
Posted on: Mon, 03/23/2009 - 09:13
Hi,
I just study your RDFS specification.
Looking at the specification of the city relation, I suspect that it does not say what it is intended to say:
<rdf:Property rdf:about="http://s.opencalais.com/1/pred/city">
<rdfs:label>city</rdfs:label>
<rdfs:comment>Name of a city, village or settlement</rdfs:comment>
<rdfs:range rdf:resource="http://s.opencalais.com/1/type/em/e/City"/>
<rdfs:domain rdf:resource="http://s.opencalais.com/1/type/em/r/CompanyLocation"/>
<rdfs:domain rdf:resource="http://s.opencalais.com/1/type/em/r/CompanyMeeting"/>
<rdfs:domain rdf:resource="http://s.opencalais.com/1/type/em/r/PersonCareer"/>
</rdf:Property>
I suppose what it should mean is that a resource that is related to another resource by the city relation is either of type CompanyLocation, or of type CompanyMeeting, or of type PersonCareer. However, the RDFS specification says:
"Where a property P has more than one rdfs:domain property, then the resources denoted by subjects of triples with predicate P are instances of all the classes stated by the rdfs:domain properties."
The same problem obviously occurs with other relations having several domain or range classes.
Best,
Uta
Trackback URL for this post:
http://www.opencalais.com/trackback/18575
I saw that you published a new version of the schema. However, this new version does not seem to address the problem of domain and range being used in a wrong way. Will there be another revision where this will be addressed? If so, when will this be available?
Best,
Uta
With the release of OpenCalais 4.3 we have published an OWL ontology. We'll be happy to hear your feedbcak.
Rafi
In the next release we'll publish an OWL ontology which will solve the problems you mentioned. As far as I know, the next release is expected towards the end of the year.
Uta, your observation is correct. As you pointed out, the object in those statements can be of one of these types. This clearly violates the RDFS semantics. We are currently working on a new revision of the RDFS where this error and others of same sort will be corrected.