Re: PROV-ISSUE-265 (TLebo): RL, why? [Ontology]

Here's an example of 'tacking on' OWL Full-level restrictions:


http://dvcs.w3.org/hg/prov/raw-file/default/ontology/ProvenanceOntologyFull.owl

(It is not complete!)

* All 'structural' subclasses have superclasses like:



    <rdf:Description rdf:about="&prov;Element">
        <rdfs:subClassOf>
            <owl:Class>
                <owl:unionOf rdf:parseType="Collection">
                    <rdf:Description rdf:about="&prov;Activity"/>
                    <rdf:Description rdf:about="&prov;Entity"/>
                </owl:unionOf>
            </owl:Class>
        </rdfs:subClassOf>
    </rdf:Description>

(ie. you can't make your own prov:Element which is not an Activity or Entity)


Adding wasInvolvedBy as inverse of prov:involved, this is used to say
for instance:

    <rdf:Description rdf:about="&prov;Derivation">
        <rdfs:subClassOf>
            <owl:Restriction>
                <owl:onProperty
rdf:resource="http://www.w3.org/ns/prov-o-full#wasInvolvedBy"/>
                <owl:someValuesFrom rdf:resource="&prov;Entity"/>
            </owl:Restriction>
        </rdfs:subClassOf>
    </rdf:Description>

ie. if :x prov:involved [ a prov:Derivation ] then :x must be an Entity.

This would solve Luc's issue with entity using entities.



However this brought up an interesting issue.. when reasoning over
this, prov:Involvement becomes a prov:Element (ie. an activity or
entity) - because it is in the range of prov:involved which also has
range prov:Element.

So Tim, what were you intending with this unified prov:involved ?  We
can easily make it work by making prov:Involvement the third Element.



On Fri, Feb 24, 2012 at 08:53, Luc Moreau <L.Moreau@ecs.soton.ac.uk> wrote:
> Hi Graham,
>
> I really invite you to go back to the minutes of F2F1 and resolution
> http://www.w3.org/2011/prov/meeting/2011-07-07#resolution_6
>
> The requirement has always been there, but has not been enforced, till
> Ivan made it clear again that it is of critical importance to adoption.
>
> We are seeing that creating a "useful" ontology without taking
> this requirement into account makes it hard to "retrofit it" later.
>
> Useful ontology, yes, but please with this requirement in mind.
>
> Regards,
> Luc
>
>
> On 02/24/2012 06:20 AM, Graham Klyne wrote:
>
> It seems to me that the RL "requirement" is a *way* lesser issue than having
> a basic model that is easy to generate.
>
> My suggestion would be push ahead with a "useful" ontology that captures a
> fair richness of provenance, and then later consider what could be traded
> off for RL compatibility.  But I do feel that expending significant effort
> on RL compatibility rather than focusing on an usefully descriptive ontology
> will probably be counter-productive at this stage.
>
> In my experience, a substantial use of OWL ontologies is for documentation
> and testing purposes, not as part of a live application (though I suppose RL
> aims to change that?).
>
> #g
> --
>
> On 24/02/2012 04:18, Provenance Working Group Issue Tracker wrote:
>
> PROV-ISSUE-265 (TLebo): RL, why? [Ontology]
>
> http://www.w3.org/2011/prov/track/issues/265
>
> Raised by: Timothy Lebo
> On product: Ontology
>
> The very recent and very hard constraint to ensure that PROV-O remains at
> OWL-RL expressivity is adding a lot of complication to the PROV-O team's
> ability to complete the ontology. VERY common restrictions that have been
> around for almost a decade and which provide a lot of insight are NOT
> permitted in OWL-RL.
>
> From what I understand, we should stay with OWL-RL so that "it stays simple
> and people will adopt it". Which people, exactly, will refuse to encode RDF
> if the corresponding ontology is more expressive than OWL-RL?
>
>
> I called in when Ivan discussed this at F2F2, and I did NOT get the
> impression that the rest of the group now seems to have. He seemed to be
> advocating for the "scruffies", which we had recently named in the meeting.
> Linked Data is the new direction for the semantic web, and that community
> could not care less about OWL expressivity. RDF and SPARQL rule the day.
> There are no OWL reasoners to be found.
>
> I consider myself a member of the Linked Data community, but I also happen
> to appreciate a well designed OWL ontology. The Linked Data community thinks
> in terms of classes and predicates. All they care about is which properties
> lead from which classes and head to which other classes. That's it. Give
> them some examples and they're off running. Oh, and make your URIs
> dereferenceable.
>
> So I don't think that the Linked Data community is going to ignore PROV-DM
> based on its OWL profile.
>
> Are we trying to satisfy some _other_ community? If so, who are they, how
> many of them are there, what do they do, and what do they like?
>
> Thanks,
> Tim
>
>
>
>
>
>
> --
> Professor Luc Moreau
> Electronics and Computer Science   tel:   +44 23 8059 4487
> University of Southampton          fax:   +44 23 8059 2865
> Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
> United Kingdom                     http://www.ecs.soton.ac.uk/~lavm



-- 
Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester

Received on Friday, 24 February 2012 16:15:15 UTC