- From: Tim Berners-Lee <timbl@w3.org>
- Date: Mon, 5 Mar 2001 18:30:48 -0500
- To: "Sigfrid Lundberg, Lub NetLab" <siglun@gungner.lub.lu.se>
- Cc: "Peter Crowther" <Peter.Crowther@melandra.com>, "pat hayes" <phayes@ai.uwf.edu>, <www-rdf-logic@w3.org>
I think we all agree that < and <= style of subProperty are both consistent
reasonable terms.
> Your view is perfectly logical, as is that x1 <= x2 && x2 <= x1 implies
> that x1 is equal to x2. However, I want subPropertyOf (X,Y) imply that Y
> has a STRICTLY NARROWER semantics than X, and the same should go for
> sameClassAs.
>
> Sigge
I can see that when building ontologies this is a typical need. i'll tell
you why
Ilike the other way.
1. Saying that subclassOf(c,d) is a way of saying forall x, in(x,c) =>
in(x,d)
which is a simple thing to say. Lots of rules systems allow that to be
expressed.
Saying properSubClassOf(c,d) is to say
forall x. in(x,c) => in(x,d) and exists y. in(x,d) and not(in(x,c))
This is a more complicated thing to say, as it uses a "not".
2. Practically, if i make an ontology I would prefer to make the first
statetemnt.
Suppose, the Robin car company asserts that is the only manufacturer
of thee-wheel cars. It does this by saying
daml:subclass(TheeWheelCar, robincar)
but given the rdfs vocabulary accidentally is led (though a gui tool) to
assert
rdfs:subclass(ThreeWheelCar, robincar)
which at the same time asserts that there is some car made by Robin which
has other than 3 wheels.
I would imagine a real case in which (say) a catlog search finds that all
robinCars
are in fact TheeWheelCars, and so the system concludes correctly that a
car is a robinCar if and only if it has three wheels. And that would be
good.
Tim
Received on Monday, 5 March 2001 18:32:43 UTC