Re: Publications about OWL (1 or 2) Full

[Strange, this whole discussion on OWL Full was captured by my SPAM 
filter - what does this mean? :->]

Am 19.05.2011 19:42, schrieb Michael F Uschold:
> Do you know of a more practice-focused
> paper that gives examples of what you can and cannot do with OWL2
> metamodelling, compared to OWL-Full?

By "can do" you certainly mean "infer", because the additional 
entailments of OWL 2 Full over OWL 2 DL are the distinctive aspect here 
(syntactically, you can write down quite a couple of things in OWL 2 DL 
that look like metamodeling without getting your parser upset). So let's 
make a few not too complicated examples for OWL 2 Full 
metamodeling-based entailments (sorry, no paper (yet)).

You can already get a few such entailments from valid OWL 2 DL 
ontologies that are themselves valid OWL 2 DL ontologies but which are 
not covered by the semantics of OWL 2 DL (the OWL 2 Direct Semantics). 
The easiest one's are those derived from "rdf:type - owl:sameAs" chains, 
such as

     (01)
     Premise:

         dbpedia:Eagle owl:sameAs freebase:Eagle .
         ex:Harry rdf:type dbpedia:Eagle .

     Conclusion (OWL Full only):

         ex:Harry rdf:type freebase:Eagle .

It's funny to note that the owl:sameAs link in the premise above really 
exists in DBPedia; the owl:type triple doesn't, though. But you can have 
a look in a previous mail that I wrote (actually, it's three mails), 
where I give a "live" example of such a chain (with countries instead of 
eagles, and England instead of Harry), and afterwards find by SPARQLing 
that there are literally thousands(!) of such chains in DBPedia and 
other LOD cloud nodes:

 
<http://lists.w3.org/Archives/Public/public-owl-dev/2010OctDec/0057.html>

But you will only experience the real power of the dark side, er, OWL 
Full, if you leave the OWL 2 DL syntactic fragment and allow to use OWL 
vocabulary, such as "rdf:type", as normal entities. It is then easily 
possible to define a "ex:metatype" property that relates individuals 
with their meta classes by

     (02) ex:metatype owl:propertyChainAxiom (rdf:type rdf:type) .

Using this with the assertions

     (03) ex:Eagle rdf:type ex:Species .
          ex:Harry rdf:type ex:Eagle .

yields

     (04) ex:Eagle ex:metatype ex:Species .

You can now encode lots of things where the three ontological layers of 
individuals, classes and meta-classes are related somehow, and get 
entailments from it. For example, you can model the following:

     "A species for which there are still member instances
      is an active species."

by

     (05) [ owl:intersectionOf (
              ex:Species
              [ owl:someValuesFrom owl:Thing ;
                owl:onProperty [owl:inverseOf rdf:type] ] )
          ]
              rdfs:subClassOf ex:ActiveSpecies .

This together with (03) gives you that eagles are (still) active 
species, because of Harry:

     (06) ex:Eagle rdf:type ex:ActiveSpecies .

What a relief! Ok, this should be enough for a start. I have many other 
examples in draft form, including much more sophisticated ones, ready to 
be written up, so you might still get your paper some day. :-)

Now, if you believe that this is all foolish, impossible and blasphemic 
stuff, just let me tell you that all my examples above are properly 
covered by the OWL 2 RL/RDF rules [1], which is a "light-weight" sub 
language of OWL 2 Full, so you don't really need a full-fledged OWL 2 
Full reasoner for processing them. And if you don't believe this, use 
Ivan Herman's online RL reasoner at

     <http://www.ivan-herman.net/Misc/2008/owlrl/>

to check (make sure to define all the additional namespace prefixes!). 
Or you can likewise use OWLIM (with the OWL 2 RL ruleset) or any other 
of the myriads of OWL 2 RL/RDF rule reasoners that are now available, 
which should give you the same results. Actually, the last example 
(05/06) is pretty much at the limit of the RL rules, and I can easily 
rephrase it in a way that it falls off from RL but still keeps within 
OWL 2 Full. But don't despair: even those harder reasoning examples can 
be managed. But that's a different story and won't be told today! :-)

Cheers,
Michael

[1] OWL 2 RL/RDF rules: 
<http://www.w3.org/TR/2009/REC-owl2-profiles-20091027/#Reasoning_in_OWL_2_RL_and_RDF_Graphs_using_Rules>

-- 
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
Stiftung Az: 14-0563.1 Regierungspräsidium Karlsruhe
Vorstand: Dipl. Wi.-Ing. Michael Flor, Prof. Dr. rer. nat. Ralf Reussner,
Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi 
Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
==============================================================================

Received on Tuesday, 24 May 2011 19:34:31 UTC