[expertfinder-dev] Re: Expertise and Trust Vocabs

[Tom Heath]

Hi Alex,

I gave this issue some more thought, and remainded unresolved about
the right approach. In hindsight even the named graphs approach didn't
seem 100% correct. So, I spoke to Chris Bizer, and his suggestion was
to use both: something like dc:date as a property of the trust
relationship (i.e. when it was generated), and another statement
(presumably using something like dc:date again) about the graph
itself, indicating when it was published. It feels a little bit messy
and like a duplication of data, but it does cover all/more
eventualities. Chris's opinion was that these sorts of issues are not
fully resolved yet, so I imagine that as more people start to use
named graphs in anger then some more discussions will be had to try
and resolve them, or at least to establish best practice.

How does this sound to you?

Cheers :)

Tom.

On 02/07/07, Alexandre Passant <alex at passant.org> wrote:
> Hi Tom,
>
> (CC-ed to the named graphs mailing-list as that's related)
>
> On 7/2/07, Tom Heath <tom.heath at gmail.com> wrote:
> > Hi Alex :)
> >
> > > > I'm not sure if you've seen my draft Hoonoh ontology Andreas. It takes
> > > > a slightly different approach, but fulfils the same kind of role. For
> > > > example, using the ExpertiseRelationship class and associated
> > > > properties one can describe an expertise relationship between a
> > > > foaf:Person and a Topic (hoonoh:Topic, subclass of skos:Concept), and
> > > > assign this relationship a value. I hope that it's a relatively clean
> > > > bit of modelling, but would be pleased to get feedback.
> >
> > > One thing that must be thinked about is also provenance of each
> > > ExpertiseRelationship statements. Once in general it is useful to have
> > > this for any RDF statements this is even more true here, when dealing
> > > with trust.
> >
> > Totally agreed.
> >
> > > Eg, I can say that X is expert in Y at 0.8 but one may think he's only 0.2.
> > > Don't we need a definedBy property ? (w/ hoonoh:ExpertiseRelationship
> > > as domain and foaf:Person as range), or the inverse, eg
> > > definesRelationship
> >
> > We'll be using Named Graphs to achieve this effect, which I think is
> > technically the right way to go and saves introducing another
> > property.
> That's also what I thaught, but I'm not sure how to manage provenance
> of relationships in a case we export all of them in a single page, eg
> [1].
> Must/Can we use swp:authority for each statement URI contained in [1]
> (see [2]) - or did I missed something with named graphs and RDF/XML ?
>
> Best,
>
> Alex.
>
> [1] http://kmi.open.ac.uk/people/tom/tmp/sample-hoonoh-instances.rdf
> [2] http://sites.wiwiss.fu-berlin.de/suhl/bizer/TriG/
>
> > Cheers,
> >
> > Tom.
> >
>