Re: ACTIO: SKOS concepts and OWL classes

Hi Guus and others,

>
> I suggest you and Ed take this text and adapt it in the way you see as 
> appropriate for the Primer.
>
> Guus

I have made a first draft at [1], Feb 5 editor's draft, section 3.3.

However, I would really like the WG that this section has quite some 
consequences on the fundamentals of SKOS.
One of the aspects I got aware of is that by allowing classes to be 
concepts or instances of classes to be concepts, we more or less deny 
what the previous SKOS guide [2] called 'indirection'.

In an oppen issues section, [2] indeed proposes that SKOS models things 
that belong to the realm of abstract concepts, while OWL classes and 
individuals can be about "concrete" things like persons. For instance, 
[2] reads:

> So, for a resource of type |skos:Concept|, any properties of that 
> resource (such as creator, date of modification, source etc.) should 
> be interpreted as /properties of a concept/, and *not* as properties 
> of some 'real world thing' that that resource may be a 
> /conceptualisation of/.

If we allow the instance of foaf:Person ex:timBL to be also a SKOS 
concept, or the class ex:Paintings (which denotes a set of painting 
objects) to be also a SKOS concept, then this indirection assumption 
seems to be at risk.
Especially, we may have "collision of properties", e.g. with the 
"creators" of a Concept.  To continue with the example of [2], a same 
object (Henry VIII) could have both Henry VII, Elizabeth of York and 
Alistair Miles as creators.

Shall we ignore this? I would feel better knowing that the issue will 
not re-surface (at least in the WG) after the documents are validated.

Best,

Antoine

[1] http://www.w3.org/2006/07/SWD/wiki/SKOS/DraftPrimer
[2] http://www.w3.org/TR/swbp-skos-core-guide#secmodellingrdf
>
>
>
>>
>> 1. When you say
>>
>>> OWL Full users will be able to handle the situation above by 
>>> treating |skos:Concept| explicitly as a metaclass, e.g. by adding 
>>> the statement of the form:
>>>     skos:Concept rdfs:subClassOf owl:Class .
>>>   
>>
>> I feel a bit unconfortable. The starting point of this SKOS/OWL issue 
>> is that we want to apply OWL modelling features to SKOS concept 
>> individuals, or SKOS modelling features to instances of OWL classes. 
>> And because this result (either because of some axiom or because of 
>> apriori specification of the type of these entities), we happen to 
>> "cross the streams", as Alistair puts it in [1], that is, being 
>> outside of OWL-DL.
>> But that's to me different from the intention of saying that every 
>> skos:Concept is an owl:Class. The statement that you write here 
>> creates the same kind of effect as the one saying that skos:broader 
>> was transitive, which we wanted to avoid because (among other 
>> reasons) it would have messed with non-transitive interpretations of 
>> broader.
>>
>> 2. Could/Should we had a sentence for the users interested in OWL-DL 
>> *and* linking classes to concepts? At the time they read the Primer 
>> this metamodelling option in OWL 1.1 could well not exist.
>> I could put in the section on "extending SKOS" a brief note on one 
>> among the patterns of [1,2] which could help them to solve their 
>> problem (e.g. the use of a bridge property such as skos:it or 
>> whatever name is convenient). And point to it from your text.
>> This would actually enable to extend this topic from "OWL classes" to 
>> "entities in the OWL ontologies", including OWL individuals such as 
>> instances of foaf:Person. This was one of the motivations for these 
>> discussions at the begining, cf. the motivation section in [2] 
>> (especially the POWDER case)
>>
>> Best,
>>
>> Antoine
>>
>> [1] 
>> http://isegserv.itd.rl.ac.uk/public/skos/2007/10/f2f/skos-owl-patterns.html 
>>
>> [2] http://www.w3.org/2006/07/SWD/wiki/SkosDesign/ConceptSemantics
>>
>>> ACTION: Alistair and Guus write draft section in primer on 
>>> relationship between SKOS concepts and OWL classes for OWL DL users 
>>> [http://www.w3.org/2008/01/22-swd-minutes.html#action06]
>>>
>>> Alistair, all,
>>>
>>> I've written a first draft. See the attached HTML.
>>>
>>> Guus
>>>
>>>
>>
>

Received on Monday, 4 February 2008 22:37:43 UTC