W3C home > Mailing lists > Public > public-cwm-talk@w3.org > July to September 2008

Fixing broken reification in OWL 2?

From: Jeff Thompson <jeff@thefirst.org>
Date: Thu, 07 Aug 2008 09:05:06 -0700
Message-ID: <489B1D32.9090502@thefirst.org>
To: Dan Connolly <connolly@w3.org>, cwm talk <public-cwm-talk@w3.org>

In N3, you can say "It's not true that Bob wrote MobyDick".
{ ex:Bob ex:wrote ex:MobyDick } a n3:falsehood .

In OWL 2 you can say this with a negative assertion:
NegativeObjectPropertyAssertion(ex:wrote ex:Bob ex:MobyDick)

As shown in http://www.w3.org/2007/OWL/wiki/Mapping_to_RDF_Graphs ,
to express this in RDF, OWL 2 introduces its own reification vocabulary.
[a owl:NegativePropertyAssertion;
  owl:sourceIndividual ex:Bob;
  owl:assertionProperty ex:wrote;
  owl:targetIndividual ex:MobyDick].

But N3 tries to correct this style of reification by quoting the URI references.
See "The quoting of URIs" in http://www.w3.org/DesignIssues/Reify.html .
The N3 reification approach would change the OWL 2 example to:
[a owl:NegativePropertyAssertion;
  owl:sourceIndividual [rei:uri "http://example.com/ex#Bob"];
  owl:assertionProperty [rei:uri "http://example.com/ex#wrote"];
  owl:targetIndividual [rei:uri "http://example.com/ex#MobyDick"]].

So my question is this: If N3 had to fix the "broken" reification in RDF,
is OWL 2 "missing the boat" to fix the issue now that it is creating
its own reification vocabulary?  Or was N3 refication trying to fix
something which is not an issue when OWL 2 is quoting an assertion in
order to negate it?

- Jeff

P.S.: Dan, I copied you because you talked about this here:
Received on Thursday, 7 August 2008 16:05:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:01:05 UTC