Re: Where DAML+OIL deviates from the RDF-Schema spec.

From: "Sigfrid Lundberg, Lub NetLab" <siglun@gungner.lub.lu.se>
Subject: Re: Where DAML+OIL deviates from the RDF-Schema spec.
Date: Mon, 5 Mar 2001 08:41:00 +0100 (MET)

> On Sun, 4 Mar 2001, Tim Berners-Lee wrote:
> 
> > DAML is in my opinion right.  If a is a subclass of b and b of a mean
> > that a and b are equivalent classes - this is not a bug.  You can't go
> > peperring the underlying logical framework with exceptions just because
> > in some cases a loop is formed by mistake. 
> 
> Tim, I agree that this isn't a bug. It is a feature, and you may or may
> not like that feature. If you're an implementor features and bugs might be
> difficult to distinguish. 
> 
> This has very little to do with underlying logical frameworks, but rather
> with the semantics of "subPropertyOf" and "subClassOf" on the one hand,
> and "samePropertyAs" and "sameClassAs", on the other. And in particular if
> samePropertyAs is a subPropertyOf subPropertyOf ;)
> 
> I belong to those that feel any set of sub-properties to a certain
> property X, should not contain any property which is equivalent with X. 
> This is like the distinction between less than or equal to '<=', or just
> less than '<'. You want subPropertyOf and subClassOf to be like '<='. But
> I want them to be like '<', for reasons similar to those put forward by
> Ian and DanBri, among others.

If you want sub... to be <, then you have to make it be <.  Right now,
there is no indication that sub... is indeed <.  In fact, there is no
problem in RDF(S) having one class be a sub of another as well as
equivalent to that other.  Adding this requirement would also be a change.

> 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. 

Go ahead and argue for this, but it is as much of a change as eliminating
the prohibition on cycles.

> Sigge

Peter Patel-Schneider

PS: Of course, a lot of this dicussion would be clearer (and cleaner) if
there was a semantics for RDF(S).

Received on Monday, 5 March 2001 08:58:08 UTC