Re: cycles in subclass hierarchy (was: RDFCore Update) (fwd)

Dan--

I agree that folks who *design* schemas with explicit subclass loops are
probably making an error of some kind.  But another use case here is
when some of the class (subclass) relationships are *derived* (e.g., the
inference mentioned in the snippet from the forwarded message below). 
For example, in Semantic Web apps I may be combining information based
on a number of different schemas, and while the individual schemas may
contain no loops, the combination may turn out to (effectively) have
one.  Rather than throwing out the resulting "effective schema" as
violating some syntactic constraint, I want to determine what having
come up with that loop "means".  I could still conclude that an error
had been made, but I might also conclude that the classes involved in
the loop were equivalent (and the designers of the individual schemas
just hadn't known about the other ones).  

--Frank

Dan Brickley wrote:
> 
> FWIW I've no objections to making the change. I still thing that folk
> designing schemas with subclass loops are making a usability error, but am
> quite happy with the idea of removing this constraint from RDFS itself.
> 
> Dan
> 
> ---------- Forwarded message ----------
> Date: Mon, 03 Sep 2001 22:56:16 +0200
> From: Frank van Harmelen <Frank.van.Harmelen@cs.vu.nl>
snip
> 
> In more detail:
> 
> 1. in DAML+OIL, subclass-relations can be inferred even if they are not
> explicitly stated (note that this is an important difference from RDF-S,
> where A is only a subclass of B if >*and only if*< it is either
> explicitly stated, or follows from subsumption in the class-hierarchy).
> 

-- 
Frank Manola                   The MITRE Corporation
202 Burlington Road, MS A345   Bedford, MA 01730-1420
mailto:fmanola@mitre.org       voice: 781-271-8147   FAX: 781-271-8752

Received on Tuesday, 4 September 2001 11:11:27 UTC