Re: question on rdfs:domain

Bijan Parsia wrote:
 > Unless I've screwed it up due to fatigue :) But I don't think so
 > glancing back.

I suspect you might need some sleep, but me too, ...

(hmmm you wrote before the weekend, me after, is work or play more 
exhausting?)


isn't the RDF/XML fragment identical in triples to

<owl:ObjectProperty rdf:ID="someProperty">
      <rdfs:domain>
        <owl:Class>
          <owl:unionOf rdf:parseType="Collection">
          </owl:unionOf>
        </owl:Class>
      </rdfs:domain>
</owl:ObjectProperty>

and hence in DL and represents a property with empty domain, i.e. an 
empty property.

Jeremy


> 
> On Feb 18, 2005, at 1:56 PM, abc def wrote:
> 
>>
>> Is the following construct valid in owl:
>>
>> <owl:ObjectProperty rdf:ID="someProperty">
>>     <rdfs:domain>
>>       <owl:Class>
>>         <owl:unionOf 
>> rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#nil"/>
>>       </owl:Class>
>>     </rdfs:domain>
>> </owl:ObjectProperty>
>>
>> The above construct is defining a property with an empty domain
> 
> 
> No, it's defining a property with a domain that consists of the class 
> rdf:nil (recently promoted from individual by  the power of OWL Full).
> 
>> which doesn't make much sense.
> 
> 
> Sure it does. It's just the inverse of a property with the empty class 
> (owl:Nothing).
> 
>>   What I am looking for is a place in the specifications (if any) that 
>> says that such a construct is valid/invalid.
> 
> 
> What sort of thing are you looking for? It's not OWL DL since the 
> collection vocabulary is forbidden in the modeling domain. But it's 
> perfectly ok, OWL Full. It really doesn't mean what you think it means, 
> but what you thought it means is also definable (even in OWL DL).
> 
> Consider the following construct (in ad hoc syntax):
>     an owl;Restriction onProperty someProperty
>                     maxCardinality = 0
> 
> and this:
> 
>     an owl:Restriction onProperty theinverseof someProperty
>                     maxCardinality = 0
> 
> and:
>     an owl;Restriction onProperty someProperty
>                     someValuesFrom owl:Nothing
> 
> and also:
>     an owl:Restriction onProperty theinverseof someProperty
>                     someValuesFrom owl:Nothing
> 
> (Hint, 1 and 3 say the same thing as do 2 and 4. If you make these a 
> superclass of owl:Thing, then 2 and 4 are the same as a empty domain.)
> 
> Unless I've screwed it up due to fatigue :) But I don't think so 
> glancing back.
> 
> I wouldn't focus on the spec per se (except to realize that owl:Nothing 
> is the empty set), but on the logic of the constructs first. If you 
> don't grasp that, the specs will not be helpful (since they, rightfully, 
> presume a fair degree of logically fluancy).
> 
> Hope this helps.
> 
> Cheers,
> Bijan Parsia.
> 
>

Received on Monday, 21 February 2005 14:17:34 UTC