W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

Re: Datatype Defaulting in OWL

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Thu, 22 Nov 2007 08:45:15 +0000
Message-Id: <9696F9CE-B13A-4482-9FA5-324B57887025@cs.man.ac.uk>
Cc: SW-forum Web <semantic-web@w3.org>, public-owl-wg Group WG <public-owl-wg@w3.org>
To: "Sean B. Palmer" <sean@miscoranda.com>

If you are really curious about the OWL perspective, public-owl-dev  
would be the right cc.

On Nov 22, 2007, at 8:27 AM, Sean B. Palmer wrote:

> Is there a way to express, in an ontology, that a property's range is
> in a particular datatype even if the datatype isn't present in the
> instance?

I don't think so. This is an RDF issue.

> In other words, say I have a property called :numberOfFish, and in one
> instance I have:
>
> :John :numberOfFish "5" .
>
> and in another:
>
> :Mary :numberOfFish "15"^^xsd:integer .
>
> In the schema I want to make it clear that lexical values must be
> taken from xsd:integer but that you're free to omit the datatype
> declaration in the instance and can consider it to be regarded an
> xsd:integer anyway. So the first triple should entail :John
> :numberOfFish "5"^^xsd:integer.
[snip]

The problem is that the "5" *does* have a type: Plain literal. Of the  
top of my head, that type is the same as xsd:string but disjoint from  
xsd:integer.

So even if you had a restriction saying that the range  
of :numberOfFish was xsd:integer, you would end up with a contradiction.

So, either one needs a kind of macro, or some sort of type coercion.

Cheers,
Bijan.
Received on Thursday, 22 November 2007 08:45:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:13:27 GMT