RE: oneOfDistinct, a proposal for 5.18

On December 20, Peter Crowther writes:
> 
> > From: Dan Connolly [mailto:connolly@w3.org] 
> > > There is an alternative, however, which I think is 
> > preferable in some ways.
> > > 
> > > OWL could have something like
> > > 
> > > <owl:AllDistinct rdf:parseType="Collection">
> > >   <owl:Person rdf:about="#John"/>
> > >   <owl:Person rdf:about="#Susan"/>
> > > </owl:AllDisjoint>
> > 
> > Yes, that's the first design that came to my mind back on 12Nov.
> > 
> > The downside of it is that it's not straightforward to use
> > in the scenario in the guide. You can't just change
> > oneOf to oneOfDistinct in...
> > 
> > <owl:Class rdf:ID="WineColor">
> >    <rdfs:subClassOf rdf:resource="#WineDescriptor"/>
> >    <owl:oneOf rdf:parseType="Collection">
> >      <owl:Thing rdf:about="#White"/>
> >      <owl:Thing rdf:about="#Rose"/>
> >      <owl:Thing rdf:about="#Red"/>
> >    </owl:oneOfDistinct>
> > </owl:Class>
> > 
> > rather, you have to write all that plus
> > 
> > <owl:AllDistinct rdf:parseType="Collection">
> >      <owl:Thing rdf:about="#White"/>
> >      <owl:Thing rdf:about="#Rose"/>
> >      <owl:Thing rdf:about="#Red"/>
> > </owl:AllDistinct>
> > 
> > I suppose that's a small price to pay versus unravelling the
> > semantics document. But I just want us to make an informed
> > decision if/when we re-open this.
> 
> Dan, I agree that it's more to type, but I think there are three useful
> upsides:
> 
> 1) It's possible to have a oneOf where some of the elements are known to
> be distinct and some aren't (because you don't state it at that point);

And the same is true the other way around - it seems quite likely that
a user would want to assert that a set of individuals are
distinct/disjoint without there ever being an occasion to use the
whole set in a description. In this case it becomes necessary to
invent meaningless axioms, such as asserting that the AllDistinct
class is a subclass of Thing, in order to make the required assertion.

Ian

> 
> 2) It separates out the assertion that these instances are distinct, so
> that a human only has to look at the assertions to find this out, rather
> than looking everywhere the instances might be used;
> 
> 3) It's possibly easier for tool developers - certainly I'd rather
> separate out the notion of distinctness from the notion of what is in
> any particular description in an OWL authoring tool.  Maybe I'm just
> biased from my OilEd use :-).  I suspect it would be easier to roundtrip
> in such a tool with the assertion separate from the description as well.
> I guess this is restating (2) but for tools.
> 
> 		- Peter

Received on Sunday, 22 December 2002 09:10:33 UTC