[foaf-dev] Advise needed for DC range definitions
mikael at nilsson.name
Fri Mar 9 09:40:31 UTC 2007
As I've mentioned before, the DCMI is in the process of assigning
domains and ranges to the Dublin Core properties  (this is a proposal
only at this stage!). We have hit upon a fundamental issue, and I would
like to hear any advise from the RDF people on this list, as it affects
the future relationship between DC and FOAF (and the rest of the SW
The issue is simple - do we want to specify ranges that are either
subclasses of owl:Thing or rdfs:Literal, but not both, or can a mixed
approach be preferable in some contexts?
Allowing ranges that can be either owl:Things or literals (such as
dcterms:description, which we are currently proposing to have an open
range of rdf:Resource) opens up for problems with OWL-DL and other
applications. However, as was recently mentioned on the SKOS list ,
the flexibility of open ranges has advantages too.
I see that FOAF has adopted a strict separated approach, where every
range is either a subclass of owl:Thing or of rdfs:Literal but not both.
Is this correct? And does the community want DC to do the same?
Thanks in advance for any comments,
<mikael at nilsson.name>
Plus ça change, plus c'est la même chose
More information about the foaf-dev