W3C home > Mailing lists > Public > public-owl-comments@w3.org > February 2009

The two imports closure definitions disagree

From: Jonathan Rees <jar@creativecommons.org>
Date: Sat, 31 Jan 2009 19:27:17 -0500
Message-Id: <C7BAF8F7-AC6C-495D-BF96-2A7AC3B3BEFA@creativecommons.org>
To: public-owl-comments@w3.org

The OWL 2 DL imports closure is subject to the prohibition on  
owl:incompatibleWith imports, while the OWL 2 Full imports closure is  
not.

It is unfortunate that there are two different definitions of imports  
closure. Ideally they should be instances of a common pattern, or one  
should be derived from the other. At least there should be some kind  
of theorem relating them, such as  map-to-SS(import-closure(map-to- 
graph(O))) ~= import-closure(O) for all (DL) ontologies O, and the  
former is OK iff the latter is.

To answer the question in the editor's note in section 3 of the RDF- 
based semantics document: The answer is yes, this document does have  
to explain what imports mean, somehow, because imports is part of the  
OWL 2 vocabulary. Maybe your worry is that imports make a lie of the  
simple statement that every RDF graph has an OWL 2 Full semantics -  
i.e. graphs (or graph sets) that are not imports closed are not given  
a meaning. I don't know how to address this correctly but you're going  
to have to come to terms with this somehow - either retract the claim  
that all graphs have a meaning, say that the meaning is not RDF- 
semantics-like (because it relies on chasing imports), or say that  
unclosed graphs/ graph sets *do* have some meaning.

Best
Jonathan
Received on Sunday, 1 February 2009 00:27:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 1 February 2009 00:27:58 GMT