W3C home > Mailing lists > Public > public-owl-wg@w3.org > March 2008

RE: Proposal to resolve ISSUE-81

From: Michael Schneider <schneid@fzi.de>
Date: Fri, 28 Mar 2008 14:31:27 +0100
Message-ID: <0EF30CAA69519C4CB91D01481AEA06A0803821@judith.fzi.de>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, <jjc@hpl.hp.com>
Cc: <public-owl-wg@w3.org>
Peter F. Patel-Schneider answered to Jeremy Carroll:

>> Negative assertions seem to me to be well out of scope of the design
>> of RDF, and unlikely to be implementable in an interioperable way in
>> typical RDF systems. 
>> 
>> I think your 'want' here is unreasonable - if you really want this
>> expressivity then take the OWL baggage. 
>> 
>> Jeremy
>
>Michael's wants appear to me to be quite reasonable.
>
>To me what he wants is that the use of negative assertions 
>only requires
>a certain kind of simple expressive power on the RDF side. This would
>allow stating something like the following
>
>	a r b
>	NOT ( a r b )
>
>in a way that is not too foreign to RDF, i.e., certainly not depending
>on semantic support for restrictions with nominals.
>
>Whether this is in the "scope of the design of RDF" is debatable.  Even
>RDF itself has contractions so one cannot argue from any claimed
>contradiction-free status for RDF.

I am still pondering about Jeremy's comment, and maybe I am on the wrong
track here...

But if contradicting property assertions such as

    a r b
    NOT ( a r b )

would really be out of scope in RDF in principle, then I would not know how
to answer the question why contradicting class assertions such as

    a rdf:type c
    a rdf:type COMPLEMENT(c)

or contradicting class axioms such as

    c1 owl:equivalentClass c2
    c1 owl:complementOf c2

should be regarded to be ok? 

At least, I don't remember having ever heard RDF people complaining about
the latter two.

But, perhaps, Jeremy's point is that "negating" RDF triples is somehow alien
to RDF. In fact, all the triples in the latter two contradicting statements
above are actually *not* of this kind, they all are "holding triples". But,
of course, the encoding which I proposed for negative property assertions
would also only consist of several "holding triples". 

But as I said, I am not certain whether I have understood Jeremy correctly.

Cheers,
Michael

--
Dipl.-Inform. Michael Schneider
FZI Forschungszentrum Informatik Karlsruhe
Abtl. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: Michael.Schneider@fzi.de
Web  : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555

FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts
Az: 14-0563.1 Regierungspräsidium Karlsruhe
Vorstand: Rüdiger Dillmann, Michael Flor, Jivka Ovtcharova, Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus


Received on Friday, 28 March 2008 13:32:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 28 March 2008 13:32:13 GMT