- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Thu, 28 Feb 2008 06:10:55 -0500 (EST)
- To: gstoil@image.ece.ntua.gr
- Cc: public-owl-wg@w3.org
From: "Giorgos Stoilos" <gstoil@image.ece.ntua.gr> Subject: RE: completeness Date: Thu, 28 Feb 2008 09:39:37 +0200 > Hello, > > See small comment inline. [...] > > > So pD*sv might be a little too much as a foundation for OWL-Prime. > > > > Probably, as I believe that rule application in pD*sv is non-terminating. > > > > Consider, for example, > > > > (1) re owl:someValuesFrom re . > > (2) re owl:onProperty p . > > (3-0) o1 rdf:type re . > > > > Then from (1,2,3-0) rule rdf-svx produces a new b-node _:b1 with > > > > (3-1) _:b1 rdf:type re . > > (4-1) o1 p _:b1 . > > > > Then from (1,2,3-1) rule rdf-svx produces a new b-node b2 with > > > > (3-2) _:b2 rdf:type re . > > (4-2) _:b1 p _:b2 . > > > > I guess you could always use some blocking condition to stop the algorithm > and then imply that > > _:b1 p _:b1 > > ...right? It is not certain that a blocking condition is possible. I do not know of any known blocking condition. Any blocking condition might have to be careful about things like owl:someValuesFrom owl:sameAs owl:allValuesFrom Note: I have not thought much about the pD* treatment of owl:sameAs as it relates to messing with the OWL vocabulary. > > And so on. This inference chain also exists for partial pD*sv > > closures. [...] peter
Received on Thursday, 28 February 2008 11:11:18 UTC