RE: ISSUE-68 (was Re: nonmon mapping and punning)

Hi, again!

I wanted to add that the problem related to ISSUE-68, which I described in my previous mail, exists for several kinds of axioms:

  * cardinality restrictions
  * sub property axioms
  * equivalent property axioms
  * functional property axioms

For example:

  O1 has only ObjectPropertyS p1,p2, and SubObjectPropertyOf(p1,p2)
  O2 has only DataPropertyS p1,p2 and SubDataPropertyOf(p1,p2)
  O3 imports O1 and O2

After RDF-mapping and merging the following RDF graph results:

  T(O*) = {
    p1 rdf:type owl:ObjectProperty .
    p1 rdf:type owl:DatatypeProperty .

    p2 rdf:type owl:ObjectProperty .
    p2 rdf:type owl:DatatypeProperty .

    p1 rdfs:subPropertyOf p2 .
  }

Again, T(O*) cannot be inverse-mapped to Functional Syntax.

The general "problem pattern" seems to be the following: An ontology O imports two (or more) other ontologies O1 and O2, where

  * O1 contains /only/ an ObjectProperty p, on which it has one of the axioms mentioned above
   (analogue argumentation for only DataProperty p ommitted here)

  * O2 contains /at least/ a DataProperty with the same name "p", and with the datatype-version of the same axiom on p. 

Even if O2 contains p both as a DataProperty and an ObjectProperty, the OWL-1.0 version of the respective axiom's representation will exist in the final merge graph, resulting from the RDF-mapping of O1 (this is actually Peter's original example). And this OWL-1.0 version of the axiom, in combination with the two different typing axioms for p, will stop the merge graph from being inverse-mapped to Functional syntax.

Best,
Michael


I wrote:

>Hi Peter!
>
>Thanks for your mail. I now see that I really did not 
>understand Jeremy's concern w.r.t. ISSUE-68 in the last telco. 
>
>When I correctly understand Jeremy, then the following variant 
>of your example comes a bit closer to what Jeremy writes in 
>his cited mail below.
>
>In OWL-1.1-DL / Functional Syntax:
>
>  * O1 has an object property p and a class c = objectatmost 5 p
>
>  * O2 has a data property p and a class c = dataatmost 5 p
>
>  * O3 imports O1 and O2
>
>Mapping these three ontologies to RDF leads to:
>
>  T(O1) = {
>    p rdf:type owl:ObjectProperty .
>    c rdf:type owl:Class .
>    c owl:equivalentClass _:r1 .
>    _:r1 rdf:type owl:Restriction .
>    _:r1 owl:onProperty p .
>    _:r1 owl:maxCardinality "5"^^xsd:nonNegativeInteger .    
>  }
>
>  T(O2) = {
>    p rdf:type owl:DatatypeProperty .
>    c rdf:type owl:Class .
>    c owl:equivalentClass _:r2 .
>    _:r2 rdf:type owl:Restriction .
>    _:r2 owl:onProperty p .
>    _:r2 owl:maxCardinality "5"^^xsd:nonNegativeInteger .    
>  }
>
>  T(O3) = {
>    [] owl:import T(O1) .
>    [] owl:import T(O2) .
>  }
>
>If T(O3) is read into a Jena OntModel (current version, not 
>OWL-1.1 aware), and it is afterwards serialized back into a 
>single RDF file, I think we receive the following:
>
>  T(O*) = {
>    p rdf:type owl:ObjectProperty .
>    p rdf:type owl:DatatypeProperty .
>
>    c rdf:type owl:Class .
>
>    c owl:equivalentClass _:r1 .
>    _:r1 rdf:type owl:Restriction .
>    _:r1 owl:onProperty p .
>    _:r1 owl:maxCardinality "5"^^xsd:nonNegativeInteger .    
>
>    c owl:equivalentClass _:r2 .
>    _:r2 rdf:type owl:Restriction .
>    _:r2 owl:onProperty p .
>    _:r2 owl:maxCardinality "5"^^xsd:nonNegativeInteger .
>  }
>
>AFAICS, The problem is now that the RDF graph T(O*) cannot be 
>mapped back to Functional Syntax. There are /two/ different 
>typing axioms for p, leading to OnlyOP(p) = false, and 
>OnlyDP(p) = false. When I correctly understand the RDF-to-Func 
>mapping described in [10], none of the mapping rules in table 
>6 can be applied.
>
>So, in the case that it is desirable to make an RDF-merged 
>ontology mappable to Functional Syntax, Jena will have to 
>perform the necessary changes to the RDF graph, or to the 
>original set of RDF graphs.
>
>@Jeremy: Is it this what you mean?
>
>Cheers,
>Michael
>
>[10] 
><http://www.w3.org/2007/OWL/wiki/Mapping_to_RDF_Graphs#Translat
>ion_from_RDF_Graphs_to_Functional-Style_Syntax>

--
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