W3C home > Mailing lists > Public > public-esw-thes@w3.org > April 2011

Re: Mapping SKOS into BFO

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Sat, 09 Apr 2011 15:51:01 +0200
Message-ID: <4DA06445.5070107@few.vu.nl>
To: public-esw-thes@w3.org
Dear Jim,


> On Thu, Apr 7, 2011 at 1:19 AM, Peter Ansell<ansell.peter@gmail.com>  wrote:
>
>> Its place may be as a superclass rather than a subclass though. If you
>
> Not necessarily. That would assert that all dependent continuents are
> concepts, which is not true from the realist perspective. The concept
> of mass is not the same as the quality mass. We could make a special
> CMO (Conceptual Model Ontology, which is what this work is for)
> Concept class, and have that be a subclass of skos:Concept and
> generically dependent continuent. But that should be a last-ditch
> effort if we can't come to a better agreement.
>
>> place skos:Concept in BFO as a subclass, you want to be sure that
>> every user will follow your extended contract, even if they produced
>> their SKOS datasets before they knew about the contract. If
>> skos:Concept is a superclass of something in BFO, then you are
>> allowing for the possibility that not all skos:Concept's will follow
>> the scientific realism definition of Concept.
>
> The contract has very few obligations. There are no property
> restrictions, for one, and the definition, as something that is not
> temporally dependent and does not have independent physical existence
> is something that seems to be inherent in the idea of what a concept
> is, regardless of the more specific definitions. If a concept is also
> being treated as a set theoretic class, then it is that too, but SKOS
> explicitly doesn't concern itself with that.


Two small cents: there's http://www.w3.org/TR/2009/NOTE-skos-primer-20090818/#secskosowl which mentions the possibility for classes to be treated as concepts as well.
Intuitively, it does not seem in blatant contraction with your proposal, though.
But this is just intuition, I must say that I don't know BFO. If you have somewhere a class for classes, and if (for some reason that wouldn't be intuitive to me) this class is asserted to be disjoint with the one of generically dependent continuents, then there might be issues.

Best,

Antoine
Received on Saturday, 9 April 2011 13:49:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 9 April 2011 13:49:44 GMT