W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2001

Re: RDFS bug "A property can have at most one range property"

From: Sean B. Palmer <sean@mysterylights.com>
Date: Fri, 16 Nov 2001 19:00:44 -0000
Message-ID: <03c901c16ed1$018ca7e0$92dc93c3@localhost>
To: <tarod@softhome.net>
Cc: <www-rdf-interest@w3.org>
Hi Marc,

>   Sorry, I think I didn't explain correctly, my english is not as
> good as I wish.

No problem; I skip through emails to quickly anyway, so it's probably
a problem on my end too :-)

> The problem I wanted to explain is why the RDFSchema
> should use the domain classes as an intersection instead of
> a union, without using DAML, ok?

Peter has already pointed out that there is an ambiguity in the exact
nature of a subject when a property is applied to it that has a domain
which is constructed from a union of two classes. Of course, this
could be resolved by inferences obtained from data elsewhere in the
store:-

   :MyArticle :date :x .
   :MyArticle :title "My Article" .
   :date rdfs:domain [ daml:unionOf (:Event :Document) ] .
   :title rdfs:domain :Article .
   :Article rdfs:subClassOf :Document .

But then, that's a disadvantage of using a union of two classes as the
object of a constraint on a property. You could still just add the
data in:-

   :Article rdf:type :Document .

But the fact is that people seem to have expected multiple ranges and
domains to have been applied conjuntively. Here's some evidence:-

According to
http://sesame.aidministrator.nl/doc/rdf-interpretation.html, SESAME
interprets multiple domains as intersections

Some OIL person talks about the topic in:
http://lists.w3.org/Archives/Public/www-rdf-comments/2000JulSep/0055

And also:-

[[[
There are a few subtle differences between domain and range
restrictions in OIL and their equivalents in RDFS. First, the
specification of OIL is very clear on multiple domain and range
restriction: these are allowed, and the semantic is the intersection
of the individual statements (conjunctive semantics). In RDFS,
multiple domain statements are allowed, but their interpretation is
the union of the classes in the statements (disjunctive semantics).
This limits the reasoning capabilities of RDFS drastically7.
]]] - http://www.cs.vu.nl/~mcaklein/papers/www10/

and as for DAML:-

[[[
It is allowed in DAML+OIL to state multiple ranges. Again, such
multiple statements must be read conjunctively: the values of the
property must satisfy all the range statements (and similarly for
multiple domain statements). Note that in this aspect, DAML+OIL
departs from the RDF Schema semantics for domain and range.
]]] - http://www.daml.org/2001/03/daml+oil-walkthru.html

[[[
what changes does DAML+OIL require in RDF(S)
multiple domains with intersection semantics
multiple ranges with intersection semantics
]]] - http://www.daml.org/2001/07/RDFS-DAML+OIL-coordination.html

Thus, all DAML compliant documents that use multiple domains and
ranges do so with the intention of expressing intersections of these
classes.

I hope that helps.

Cheers,

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Friday, 16 November 2001 14:00:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:52 GMT