[Pellet-users] dateTime reasoning

[Michael Smith]

On Thu, 2007-05-10 at 23:19 +0200, sjtirtha wrote:

>                 <owl:allValuesFrom>
>                     <rdf:Description>
>                         <rdf:type rdf:resource="&owl;DataRange"/>
>                         <owl11:minInclusive
> rdf:datatype="&xsd;dateTime">2006-01-01T12:00:00</owl11:minInclusive> 
>                         <owl11:maxInclusive
> rdf:datatype="&xsd;dateTime">2007-05-31T12:00:00</owl11:maxInclusive>
>                         <owl11:onDataRange
> rdf:resource="&xsd;dateTime"/> 
>                     </rdf:Description>
>                 </owl:allValuesFrom>

>                 <owl:allValuesFrom>
>                     <rdf:Description> 
>                         <rdf:type rdf:resource="&owl;DataRange"/>
>                         <owl11:minInclusive
> rdf:datatype="&xsd;dateTime">2005-01-01T12:00:00</owl11:minInclusive> 
>                         <owl11:maxInclusive
> rdf:datatype="&xsd;dateTime">2005-05-31T12:00:00</owl11:maxInclusive>
>                         <owl11:onDataRange
> rdf:resource="&xsd;dateTime"/> 
>                     </rdf:Description>
>                 </owl:allValuesFrom>

> should not have intersection.
> But Pellet doesn't identify, that there is no intersection. 

I've tested this and believe that Pellet 1.4 behaves correctly with the
intersection of these descriptions.  Note that because they are both
universal quantifications (using allValuesFrom), the intersection is
satisfiable.  It is the case that no instances in the intersection can
exist for which the validDate property is used.  Changing either
restriction to be an existential (someValuesFrom) makes the class
unsatisfiable.

Regards,
-- 
Mike Smith

Clark & Parsia