Re: Cardinality restrictions (was: comparing DAML-ONT and OIL)

Frank van Harmelen wrote:
[...]
> One of the things that struck me in Peter's list of comparing OIL-Standard and DAML-ONT was:
> 
> >                                 OIL-Standard            DAML-1.2
> >
> >   Cardinality Restrictions      local (and global)      global only
> 
> This means that in DAML-ONT I cannot have a property "hasWheels" with a different cardinality restriction for bicycles and cars (since only one global cardinality restriction is allowed per slot).
> 
> All other slot-restrictions in DAML-ONT are local (as they are in OIL).

No, daml:domain and daml:range are global. Well.. there's a local
version of range. A local version of domain has
been suggested, I think.

> Why has cardinality been handled differently?

No particular reason... the local (i.e. per-domain)
parts of DAML were added sort of in a hurry, and
we didn't really finish the job.

Since
	-- these requests would make DAML consistent with OIL, and
	-- they have been made and seconded, and
	-- nobody has objected (including me), but
	-- I'm not in a good position to edit it into the spec
		just now
I consider these TODO items. i.e. range-dependent-domain
and domain-dependent-cardinality.
(along with defined classes; cf my message of
Thu, 12 Oct 2000 09:02:59 -0500
http://lists.w3.org/Archives/Public/www-rdf-logic/2000Oct/0019.html

though it's starting to look like the need for all of
OIL's defined classes may be met by the current DAML draft.)

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Saturday, 14 October 2000 23:26:39 UTC