W3C home > Mailing lists > Public > semantic-web@w3.org > July 2010

RE: Subjects as Literals, [was Re: The Ordered List Ontology]

From: Michael Schneider <schneid@fzi.de>
Date: Thu, 1 Jul 2010 13:54:45 +0200
Message-ID: <0EF30CAA69519C4CB91D01481AEA06A001F07AE8@judith.fzi.de>
To: <nathan@webr3.org>
Cc: "Linked Data community" <public-lod@w3.org>, "Semantic Web" <semantic-web@w3.org>
Nathan wrote:

>re OWL DL, does it have to consider every triple in a 'graph'?

No, and it cannot do so in general. Strictly speaking, OWL DL doesn't even
have a notion of RDF triples or RDF graphs. OWL DL "thinks" in terms of
constructs such as axioms and class expressions. The genuine "abstract"
syntax of OWL (2) DL, defined in [1], deals with such constructs, not with
RDF triples. OWL DL rather uses RDF as a serialization syntax, via the RDF
mapping defined in [2]. 

Basically, for each of the constructs in the abstract OWL DL syntax there is
a set of RDF triples encoding this construct. Sets of RDF triples that do
not match the serialization of any of the OWL DL constructs (or of sets of
them) are simply not syntactically valid and have no semantic meaning in OWL
DL. In fact, the whole RDF graph will be regarded as an invalid OWL DL
ontology if such sets of triples exist.

As an example, consider the RDF serialization of the OWL DL construct for
universal property restrictions:

       _:x rdf:type owl:Restriction .
       _:x owl:onProperty ex:p .
       _:x owl:allValuesFrom ex:c .

When this RDF serialization is used as the input to an OWL DL system, the
system will try to map it into the corresponding "abstract" OWL DL syntax
construct. This is only a one-or-nothing mapping, i.e. a single RDF triple
from the triple set above has no meaning for OWL DL. For instance, if the
OWL DL system only finds the triple

    _:x owl:allValuesFrom a:Dog .

and not the other two triples, then the RDF encoding of the universal
property restriction is simply broken, and you will probably receive an
error. It is /not/ the case, as one might think, that this single triple is
still sort of a property assertion with the property owl:allValuesFrom. 

Note that you have a triple-focused view in RDF(S) and OWL Full, where the
single triple is regarded as an axiom on its own. But, as stated, OWL DL
requires that the set of all RDF triples is the encoding of some set of OWL
DL constructs, otherwise the RDF is just something that is not well-defined.


Michael

[1] <http://www.w3.org/TR/2009/REC-owl2-syntax-20091027>
[2] <http://www.w3.org/TR/2009/REC-owl2-mapping-to-rdf-20091027/>

--
Dipl.-Inform. Michael Schneider
Research Scientist, Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: michael.schneider@fzi.de
WWW  : http://www.fzi.de/michael.schneider
=======================================================================
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, RP Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
=======================================================================
Received on Thursday, 1 July 2010 11:55:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:20 UTC