RE: RE: Recommendations: URIs

I'm pretty sure this does force it into OWL Full. I can imagine ways to
support both rdfs:Property and skos:Concept types in OWL DL, but it
would require more URIs and linking inside the ontology. Assuming that
OWL reasonsers become more powerful and popular, the problems with OWL
Full ontologies will presumably become more noticeable. It's not clear
how much we would sorry about such things today. It might be nice to
take an example like this and imagine how it could be migrated from OWL
Full to OWL DL without causing too much disruption.

Jeff

> -----Original Message-----
> From: Svensson, Lars [mailto:L.Svensson@dnb.de]
> Sent: Thursday, April 28, 2011 6:15 AM
> To: Young,Jeff (OR); Karen Coyle; public-lld@w3.org
> Subject: AW: RE: Recommendations: URIs
> 
> Jeff wrote:
> >
> > Karen,
> >
> > I believe that in OWL, individuals are considered to be part of the
> > ontology. Here's some wording in the OWL guide:
> >
> > http://www.w3.org/TR/2004/REC-owl-guide-20040210/#SimpleClasses
> > http://www.w3.org/TR/2004/REC-owl-guide-20040210/#Glossary
> >
> > TBox and ABox are the only terms I've come across that make the
> > distinction. In practice, though, modeler's have the ability to blur
> > the
> > line or possibly even switch a term between TBox/ABox as their model
> > evolves. For example, MARC relator codes are defined as both
rdf:type
> > skos:Concepts (ABox) and rdf:Property (TBox).
> 
> Doesn' that automatically make it OWL Full?
> 
> //Lars
>

Received on Thursday, 28 April 2011 14:46:47 UTC