W3C home > Mailing lists > Public > www-rdf-interest@w3.org > October 2000

Re: OIL Class Expressions in RDF

From: Ian Horrocks <horrocks@cs.man.ac.uk>
Date: Tue, 10 Oct 2000 19:01:39 +0100 (BST)
To: Tom Van Eetvelde <tom.van_eetvelde@alcatel.be>
Cc: www-rdf-interest@w3.org, Jeen Broekstra <jbroeks@cs.vu.nl>
Message-ID: <14819.21887.215382.734139@galahad.cs.man.ac.uk>
Hi,

I hope you don't mind me muscling in with a few answers:

On October 10, Tom Van Eetvelde writes:
> Hello Jeen,
> 
> I was not able to study your document right away due to lack of time. Anyway, I have started now
> :-).
> 
> Reaching page 5, I already have some questions. Not really about OIL, but about the example
> material. I drew a picture showing the subclass of hierarchy (see attachment). Here are my
> questions:
> 
>    * Why would one like to state that plant is not a subclass of animal? Isn't it sufficient to
>      leave out any relationship with animal? Then according to the model, these two are not related.
>      I can see some sense in stating that plant is everything that is not an animal. But then, the
>      modelling should not use 'subclassof' I guess.

Saying that a plant is a subclassof (NOT animal) is not the same as
saying that a plant is not a subclass of animal. In the former case,
we forbid any individual to be both a plant and an animal (which is
exactly what we intend) - it is the same as saying that plant and
animal are disjoint. In the latter case we are simply stating that
there is at least one plant that is not an animal.

>    * Isn't the lion misplaced?

Sort of. It is implicit from the description of lion that it is a kind
of carnivore. Intelligent OIL tools are able to find this inference
(and much more complex ones), and we will soon be releasing an OIL
editor that can automatically enrich the RDFS with explicit subclassof
relations derived from such inferences.

>    * I have the feeling that carnivore and 'not carnivore' should be on the same 'level' in the
>      hierarchy. However, 'not carnivore' is one hop away from herbivore and 'carnivore' is 2 hops
>      away. This implies a difference in the hierarchy tree, or that 'not carnivore' has a higher
>      abstraction level than 'carnivore'. Or am I seeing things the wrong way here?

Things are a bit strange with these negated concepts. Perhaps if you
think of the statement about a plant not being a carnivore as a
"disjoint" relation between plant and animal then the picture becomes
a little clearer. In the latest version of OIL we have explicit
disjointness statements and the example should be changed to reflect
this.

>    * When I say that class a is sublcassof b or subclassof c, what can I do with this information?
>      Why would you want to model classes via an 'OR'?

The example says it all - herbivores don't just eat plants, they eat
"plants or parts-of plants". If we didn't model it this way then we
wouldn't discover that a giraffe is a herbivore because leaves are not
kinds of plants they are parts of plants.

> 
> That's it for now :-).
> 
> regards,
> 
> Tom.
> <!doctype html public "-//w3c//dtd html 4.0 transitional//en">
> <html>
> Hello Jeen,
> <p>I was not able to study your document right away due to lack of time.
> Anyway, I have started now :-).
> <p>Reaching page 5, I already have some questions. Not really about OIL,
> but about the example material. I drew a picture showing the subclass of
> hierarchy (see attachment). Here are my questions:
> <ul>
> <li>
> Why would one like to state that plant is not a subclass of animal? Isn't
> it sufficient to leave out any relationship with animal? Then according
> to the model, these two are not related. I can see some sense in stating
> that plant is everything that is not an animal. But then, the modelling
> should not use 'subclassof' I guess.</li>
> 
> <li>
> Isn't the lion misplaced?</li>
> 
> <li>
> I have the feeling that carnivore and 'not carnivore' should be on the
> same 'level' in the hierarchy. However, 'not carnivore' is one hop away
> from herbivore and 'carnivore' is 2 hops away. This implies a difference
> in the hierarchy tree, or that 'not carnivore' has a higher abstraction
> level than 'carnivore'. Or am I seeing things the wrong way here?</li>
> 
> <li>
> When I say that class a is sublcassof b or subclassof c, what can I do
> with this information? Why would you want to model classes via an 'OR'?</li>
> </ul>
> That's it for now :-).
> <p>regards,
> <p>Tom.</html>
> begin:vcard 
> n:Van Eetvelde;Tom
> x-mozilla-html:FALSE
> org:Alcatel Bell
> adr:;;Francis Wellesplein 1;2018 Antwerp;;;Belgium
> version:2.1
> email;internet:tom.van_eetvelde@alcatel.be
> title:Research Engineer
> x-mozilla-cpt:;0
> tel;work:32 (0) 3 240 4181
> fn:Tom Van Eetvelde
> end:vcard
Received on Tuesday, 10 October 2000 13:49:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:44 GMT