- From: Alan Ruttenberg <alanruttenberg@gmail.com>
- Date: Tue, 1 Jul 2008 12:26:02 -0400
- To: Evan Wallace <ewallace@cme.nist.gov>
- Cc: public-owl-wg@w3.org
OK, it's clear now that there are legitimate ways in which one could consider properties and classes to denote, in some way, the same thing. What I'm missing is the purpose of the punning or put another way, the intended semantics. In the case of class/instance punning there are two things I am aware of. 1) The OWL DL "missing entailment" x sameAs y => x equivalentClass y 2) The use of the pun for the purpose of making queries, in, e.g. sparql. Are there analogous situations for the class/property? Could the concrete benefit of punning in this case be made explicit? Is it a matter of documentation? Should there be an entailment x equivalentProperty y <=> x equivalentClass y? What becomes easier if the property and class are named the same in such cases? Evan, would it be possible to share a model which uses this construct so we can see how it plays out? Regards, Alan On Jul 1, 2008, at 12:02 PM, Evan Wallace wrote: > > > As I said when we discussed this in January, combining properties > and classes is common in > some models. UML supports this with the Association Class > construct, and I saw this construct > used liberally in models at last week's OMG meeting. It is also > true that it is common to encounter > the problem of integrating models after the fact that have made > different design choices leading > to a need to reconcile the use of both classes and properties for > the same concept. As long as > we are going to include punning in the OWL2 language, then we should > have a darn good reason if > we are not going to support class-property punning. > > -Evan > >
Received on Tuesday, 1 July 2008 16:26:39 UTC