- From: Natasha Noy <noy@SMI.Stanford.EDU>
- Date: Tue, 17 Aug 2004 10:10:15 -0700
- To: Eric Jain <Eric.Jain@isb-sib.ch>
- Cc: public-swbp-wg@w3.org
> Natasha Noy wrote:
>> There is nothing wrong with having classes in the hierarchy be
>> instances of some metaclasses (in fact, i would argue very strongly
>> that there are many cases where you would want just that). However,
>> in your particular example, you want to represent a specific lion in
>> the San Diego Zoo as a class, and that's a problem. A class of what?
>> What are instances of this class? etc.
>
> I think I see what you mean. If we are looking at a lion, we know that
> we can't be any more specific. Therefore classes should not be used
> for things that can be pointed to. (But inversely, things that do not
> exist may still be considered to be things...)
>
> Incidently, how do you best translate rdf:type and rdfs:subClassOf
> into natural language without loosing the distinction between the two
> terms? ("is a" seems to be used for both...)
The explanation that I often found useful is to think of classes as
sets of their instances (so, the class of Lions is a set of all lions).
Then subclass-of is a subset relationship (the set of African lions is
a subset of a set of all lions). instance-of is set membership: Simba
is a member of the set of African lions (and lions, of course).
Natasha
Received on Tuesday, 17 August 2004 17:10:16 UTC