- From: Jeff Z. Pan <jpan@csd.abdn.ac.uk>
- Date: Wed, 06 Dec 2006 09:25:05 +0000
- To: Holger Knublauch <holger@topquadrant.com>
- CC: public-owl-dev@w3.org
Hi Holger, > > I am wondering why user-defined datatypes are not modeled as instances > of the RDF Schema class rdfs:Datatype (similar to the hack suggested > in the Protege 3 implementation [1]). Without knowing the design > decisions that lead to the use of owl:DataRange, my naive point of > view would be that rdfs:Datatypes may make it more consistent with the > semantic web stack. I am sure the working group had good reasons for > selecting owl:DataRange, but it would be useful to understand them > from the outside. > In RDF, a datatype (instance of rdfs:Datatype) and a class (instance of rdfs:Class) can share instances, such as the integer 42. While in OWL DL (and hence OWL 1.1), the datatype domain is disjoint with the object domain; therefore, sharing instances is not possible for an OWL DL/1.1 datatype and an OWL DL/1.1 class. Greetings, Jeff > Also, I think we should use the xsd namespace for the facet names, so > that they are written as xsd:minInclusive. > > Could anyone please clarify these issues? > > Thanks > Holger > > PS: The family.owl linked from the OWL 1.1 web site currently appears > to be inconsistent with the RDF mapping spec (at least with respect to > the user-defined datatypes). > > [1] http://protege.stanford.edu/plugins/owl/xsp.html > -- Dr. Jeff Z. Pan (http://www.csd.abdn.ac.uk/~jpan/) Department of Computing Science, The University of Aberdeen
Received on Wednesday, 6 December 2006 09:28:48 UTC