W3C home > Mailing lists > Public > www-webont-wg@w3.org > February 2003

imports and entailments

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Mon, 03 Feb 2003 18:20:04 +0000
Message-ID: <3E3EB2D4.8040305@hpl.hp.com>
To: www-webont-wg <www-webont-wg@w3.org>


Another issue that we may need to look at ...

The correspondence theorem excludes imports - are there any interesting 
entailments that behave differently.

e.g.

FileA1:

<eg:a> rdf:type owl:Class .


FileB1:

<FileB1> owl:import <FileA1>.



Does FileA1 entail FileB1?
(I believe that both rdfs compatible and direct currently say yes)

FileA2:

<FileA2> rdf:type owl:Ontology .
<eg:a> rdf:type owl:Class .


FileB2:

<FileB2> rdf:type owl:Ontology .
<FileB2> owl:import <FileA2>.


Does FileA2 entail FileB2?
(I believe that direct says yes, and rdfs-compatible says no)


Of course its gets worse if we annotate the ontology FileB2

e.g.

FileB3:
<FileB3> rdf:type owl:Ontology .
<FileB3> rdfs:comment "This is the same as FileA2" .
<FileB3> owl:import <FileA2>.


Does FileA2 entail FileB3?
(I believe that direct says yes, and rdfs-compatible says no,
but this is partly overlap with the orthogonal issue to do with 
annotations).


====

Of course, imports syntax is also a bit broke at the moment, in that the 
imports closure of FileB2 and FileB3 do not correspond to any abstract 
ontology, because they include the two triples

<FileA2> rdf:type owl:Ontology .
<FileB2> rdf:type owl:Ontology .

and the abstract syntax plus mapping rules only permit one triple of 
such a form.
I assume that will get fixed sooner or later.
(I am working on a much more extensive fix to syntactic problems, so I 
am not requesting any syntactic changes now)

Jeremy
Received on Monday, 3 February 2003 13:15:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:57 GMT