W3C home > Mailing lists > Public > public-swbp-wg@w3.org > August 2004

Re: [ANN] Working drafts by SW Best Practices group -- request for comments

From: Natasha Noy <noy@SMI.Stanford.EDU>
Date: Tue, 10 Aug 2004 12:53:10 -0700
Message-Id: <E7E7F872-EB06-11D8-AEA1-000A958B5C28@smi.stanford.edu>
Cc: public-swbp-wg@w3.org
To: Eric Jain <ejain@isb-sib.ch>

>>>>> :AfricanLion
>>>>>       a       :Animal;
>>>>>       rdfs:subClassOf :Lion .
>>>>> :Animal
>>>>>       a       owl:Class;
>>>>>       rdfs:subClassOf owl:Class .
>> AfricanLion will also be a subclass of owl:Class, which will mean that
>> all its instances are classes. In most applications, this is probably
>> not what you want.
> I realize that this approach can be misleading. Let's say we have
> :hasAnimal
>   a :ObjectProperty
>   rdfs:range :Animal
> and
> :Simba
>   rdfs:subClassOf :AfricanLion;
> then we can't say
> :SanDiegoZoo
>   :hasAnimal :Simba
> until we also assert that
> :Simba
>   a :Animal;
> Also, this will only work with OWL Full. Neverthless I think this is a
> valid approach (let me know if it isn't), and no more awkward than 
> some of
> the other solutions you propose :-)

Actually, I think it is more awkward. Having Simba both as an instance 
and a subclass of Animal is awkward at best. In some, very rare, 
circumstances, you could argue that that's what you want (in fact, I've 
done that myself [1]), but it complicates things significantly, and 
probably shouldn't be done unless you absolutely have to.


[1] N. Noy, M. Musen, J. L.V. Mejino, Jr., and C. Rosse "Pushing the 
Envelope: Challenges in a Frame-Based Representation of Human Anatomy." 
Data and Knowledge Engineering Journal,  48/3 pp. 335-359. Available as 
SMI technical report SMI-2002-0925 at
Received on Tuesday, 10 August 2004 19:53:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:17:12 GMT