- From: Alan Ruttenberg <alanruttenberg@gmail.com>
- Date: Tue, 22 Apr 2008 12:03:27 -0400
- To: Uli Sattler <sattler@cs.man.ac.uk>
- Cc: Bijan Parsia <bparsia@cs.man.ac.uk>, OWL Working Group WG <public-owl-wg@w3.org>
Thanks Bijan, Uli.
My use case would be satisfied with the simplest of keys - no
inferred keys and keys only on named classes.
-Alan
On Apr 22, 2008, at 5:56 AM, Uli Sattler wrote:
>
> On 22 Apr 2008, at 10:37, Bijan Parsia wrote:
>>
>> On 22 Apr 2008, at 10:15, Alan Ruttenberg wrote:
>>> On Apr 22, 2008, at 4:46 AM, Bijan Parsia wrote:
>>>>
>>>> Disjunction and class expressions leap to mind as things that
>>>> would need consideration.
>>>
>>> Do you have a quick example to get me thinking about this?
>>
>> It all depends on how keys get on classes. If we are punning (and
>> thus can infer keys), then just make your class and instance of
>> having key1 or key2. Similarly, can keys be asserted on anonymous
>> classes, i.e., class expressions? What does that mean?
>>
>
> here is another example:
>
> ClassKey keyPropX
>
> subClassOf ClassA SomeValuesFrom keyPropX OneOf("17"^^xsd:integer
> "18"^^xsd:integer)
>
> subClassOf ClassB SomeValuesFrom keyPropX OneOf("17"^^xsd:integer
> "18"^^xsd:integer)
>
> subClassOf ClassC SomeValuesFrom keyPropX OneOf("17"^^xsd:integer
> "18"^^xsd:integer)
>
> This implies that either
> ClassA and ClassB are identical and thus equivalent, or that
> ClassB and ClassC are identical and thus equivalent, or that
> ClassA and ClassC are identical and thus equivalent...now we have
> have real trouble with these since we cannot handle disjunctions of
> equivalences easily...
>
> Cheers, Uli
>
>> Cheers,
>> Bijan.
>>
>>
>
Received on Tuesday, 22 April 2008 16:04:13 UTC