Re: Relationships involving collections

Leonard Will wrote:
>
> On Tue, 5 Feb 2008 at 13:50:24, Jakob Voss <jakob.voss@gbv.de> wrote
>>
>> Leonard Will wrote:
>>
>>> You have shown "Gamma rays" as "Collection", but I hope that nobody
>>> would interpret it as that, in SKOS or elsewhere. It looks like an
>>> ordinary concept, and doesn't become a "collection" just because it has
>>> some narrower terms. As I understand it, "collection" is the SKOS term
>>> for what the thesaurus standard BS8723-1 calls an "array": a group of
>>> sibling terms. An array may be preceded by a node label showing a
>>> "characteristic of division", e.g. (... by wavelength) in the above
>>> example. Node labels are just interpolated explanatory labels; they do
>>> not represent concepts and they do not have relationships.
>>
>> I don't see the problem: Of course you can create an array named 
>> "gamma rays" it makes a lot sense.
>
> "Naming" an array is not something recognised in the British Standards 
> for thesaurus construction. "Gamma rays" is the label for a concept, 
> which may have a set of narrower concepts which we call an array, but 
> "gamma rays" is not the name of the array, it is the label of the 
> parent concept.
The problem for us is that we are trying to generate a skos version of 
someone else's vocabulary which should not alter the structure or the 
available terms. First, we should note that there is not a "Gamma rays" 
term in the A&A vocabulary, however, there is a collection of three 
terms that involve gamma rays. As such, the most natural mapping 
declaration for a user who does not understand vocabulary issues is to 
say that the gamma ray collection is an exact match for the gamma ray 
concept in the AOIM vocabulary. When this is translated to skos we now 
believe that each member of the gamma ray collection in the A&A 
vocabulary should be a skos:narrowMatch for the gamma ray concept in the 
AOIM vocabulary, i.e.

    aoim:gammaRay skos:narrowMatch aAndA:gammaRayBursts,                
                aAndA:gammaRayObservations, aAndA:gammaRayTheory .

We are going to go ahead with Jakob's suggestion of implementing default 
rules with user feedback for confirmation of the inferred mappings.

Thanks for all the help,

Alasdair