[Pellet-users] Unexplained Pellet Consistency Check

[JPQ]

Hi, folks.

I posted this question in the TopBraid Composer forum but was asked to post
it here. I'd appreciate your feedback. Thanks.

Consider the following simple case.

1. An "integer_value" datatype property; no domain/range assigned.

2. A "HoldOneInteger" class
           <owl:Class rdf:ID="HoldOneInteger">
             <owl:equivalentClass>
               <owl:Restriction>
                 <owl:cardinality rdf:datatype="
http://www.w3.org/2001/XMLSchema#int"
                 >1</owl:cardinality>
                 <owl:onProperty rdf:resource="#integer_value"/>
               </owl:Restriction>
             </owl:equivalentClass>
             <owl:equivalentClass>
               <owl:Restriction>
                 <owl:allValuesFrom rdf:resource="
http://www.w3.org/2001/XMLSchema#integer"/>
                 <owl:onProperty rdf:resource="#integer_value"/>
               </owl:Restriction>
             </owl:equivalentClass>
           </owl:Class>

3. A "SingleInteger" individual on the "HoldOneInteger" class

The Resource Form for "SingleInteger" displays and
unassigned "integer_value" property beneath "Other Properties", as expected.


1. Assign a value to the property. Run Pellet consistency check. Passes.
Expected.

2. Assign two values to the property. Run Pellet consistency check. Fails.
Expected - violates cardinality constraint in property restriction on
"integer_value".

3. Do not assign a value to the property. Run Pellet consistency check.
Passes. Unexpected.

Perhaps someone can clarify what occurs with unassigned values. Why doesn't
a Pellet consistency check violate the cardinality constraint, which states
there must be exactly one integer value assigned to the "integer_value"
property for an individual of type "HoldOneInteger". My only assumption is
that Pellet interprets an unassigned property value as an empty value. In
this case the set of allowable values is any <xs:integer>, and like any
other set the empty set is a valid subset.


Anthony
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.owldl.com/pipermail/pellet-users/attachments/20080602/e993399e/attachment.htm