Re: loop free?

I still don't see any particular rationale to forbid loops in
equivalentClass, as loops there as semantically vacuous.  The situation is
different with respect to disjointWith, as a loop there makes a fairly
strong statement.

Peter F. Patel-Schneider


From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Subject: Re: loop free?
Date: Tue, 30 Sep 2003 21:46:42 +0100

> 
> A loop is a technical term in graph theory for a cycle of length of one.
> 
> I personally find that the current text prohibits that with the words 
> "pair" and "undirected graph"  (most authors permit loops in directed 
> graphs and prohibit them in undirected graphs).
> 
> If we have a test case that includes a loop and is hence not in DL then 
> that will be adequately clear, and we can avoid textual change to S&AS.
> 
>  From the design point of view since the disjointWith does not allow loops, 
> it seems slightly cleaner to prohibit them for equivalentClass - it's 
> easier to code though without such a constraint.
> 
> Jeremy
> 
> 
> Peter F. Patel-Schneider wrote:
> 
> > I don't particularly care one way of the other here, but I don't see any
> > need to forbid loops (of any sort).  As well, the wording that you propose
> > could be read as forbidding non-trivial loops.
> > 
> > peter
> > 
> > 
> > 
> > From: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
> > Subject: loop free?
> > Date: Tue, 30 Sep 2003 15:19:57 +0200
> > 
> > 
> >>Concerning
> >>http://www.w3.org/TR/2003/CR-owl-semantics-20030818/
> >>
> >>
> >>Peter,
> >>
> >>an editorial suggestion on the mapping rules:
> >>
> >>[[
> >>T(descriptioni) owl:equivalentClass T(descriptionj) .
> >>for all <i,j> in G where G is a set of pairs over {1,...,n}x{1,...,n}
> >>that if interpreted as an undirected graph forms a connected graph for
> >>{1,...,n}
> >>]]
> >>
> >>suggest
> >>
> >>s/a connected graph/a loop-free connected graph/
> >>
> >>I think the case <i, i> in G is already excluded by the word "pairs" but it
> >>is arguable. For most readers undirected graphs are loop free by definition;
> >>but since we do not provide one ...
> >>
> >>
> >>
> >>
> >>
> >>(I should add a test case for
> >>
> >>_:b owl:equivalentClass _:b .
> >>_:b rdf:type owl:Class .
> >>_:b owl:unionOf rdf:nil .
> >>
> >>as being in OWL Full, similarly
> >>
> >>_:b owl:disjointWith _:b .
> >>_:b rdf:type owl:Class .
> >>_:b owl:unionOf rdf:nil .
> >>)
> >>
> >>My code, which now passes all the tests except I5.8-016, would fail those
> >>two I think :(
> >>
> >>Jeremy
> >>
> >>
> >>
>

Received on Tuesday, 30 September 2003 17:09:07 UTC