Re: Mapping SKOS into BFO

On 18 April 2011 17:45, Jim McCusker <james.mccusker@yale.edu> wrote:
> On Mon, Apr 18, 2011 at 10:34 AM, Dan Brickley <danbri@danbri.org> wrote:
>
>> IMHO foaf:focus works well for specific named entities; it works less
>> well for classes, and I have been wondering how best to address this.
>>
>> The difference is in the pragmatics of whether we say that foaf:focus
>> is a functional property. If the SKOS concept under consideration is
>
> I agree. I couldn't find anything that asserted that foaf:focus is a
> functional property, but then FOAF is RDFS.

We ought to make the functional / inverse functional flags more
visible in the spec, I guess.

FOAF has used DAML+OIL and later OWL alongside RDFS, so if you check
in the RDF version of the spec,
e.g. (not very semantically, but hey :) ... with
curl xmlns.com/foaf/spec/index.rdf | grep Functional
...you'll see a bunch of other properties are indicated with
functional or inverse-functional from OWL.

>> #fido_the_dog, then its foaf:focus would be Fido, a Dog. And if there
>> were two or three URIs around (dbpedia, freebase, VIAF, ...) for that
>> self-same entity, they're all owl:sameAs each other. If the SKOS
>> concept were #dogs, ... we could still use foaf:focus to point to
>> (various different) classes corresponding to the class of things that
>> are dogs. But having an implied sameAs amongst them all is likely to
>> be less useful, less accurate, and more contentious. Since it is
>> tempting to declare foaf:focus functional, this would likely mean
>> nudging out the class use case to a companion property, eg.
>> focusClass. Or maybe there's an OWL2 idiom that can accomodate having
>> it both ways...
>
> The functionality of foaf:focus would not affect the actual class,
> just the pun of the class. So if there were two classes that were
> represented by the same concept, the puns would be declared sameAs.
> This could affect any other metamodeling, but not the OWL classes
> themselves.

Yup, I can imagine implementors in the OWL-Full tradition doing things
a little differently. Either way, it's fiddly...

> On the other hand, cmo:represents is not functional, and would
> therefore play nicer with existing metamodels. It also leaves room for
> self-representation (x cmo:represents x, which implies that x is a
> purely conceptual thing), which, if foaf:focus were defined as
> functional, then if it were found that there is a real thing (class,
> property, or instance) that is also represented by that thing, the
> concept and the thing would be identified as sameAs.

I'm reminded dimly of UMBEL here - http://umbel.org/vocabulary - any relation?

Dan

> Jim
> --
> Jim McCusker
> Programmer Analyst
> Krauthammer Lab, Pathology Informatics
> Yale School of Medicine
> james.mccusker@yale.edu | (203) 785-6330
> http://krauthammerlab.med.yale.edu
>
> PhD Student
> Tetherless World Constellation
> Rensselaer Polytechnic Institute
> mccusj@cs.rpi.edu
> http://tw.rpi.edu
>

Received on Monday, 18 April 2011 15:57:54 UTC