Semantics Wrinkle

(continuing thread "Semantics Review")
e.g.

http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0337.html

I have been thinking some more about the text:
[[
Individual-valued properties that are transitive, or that have transitive 
sub-properties, may not have cardinality conditions expressed on them, either 
in restrictions or by being functional, or inverse functional. This is needed 
to maintain the decidability of the language. 
]]

I now believe that this wrinkle does need ironing out. (i.e. the constraint 
expressing more clearly).

The test case I find compelling is:


FunctionalProperty( <eg:f> )
EnumeratedClass( <eg:single>, <eg:a> )
Individual( <eg:a> )
ObjectProperty( <eg:p>, domain( <eg:single> ), range( <eg:single> ) )

entails (according to the 2nd Jan semantics, but shouldn't)

Individual(<eg:a>, value( <eg:p> <eg:a> ) )


Why?
<eg:p> being a property on a singleton set is necessarily transitive.
By the questionable text (deliberately misread?) it is not a sub-property of 
<eg:f>. Therefore it is not the empty property, therefore it must hold on 
<eg:a>.

Broken. (Slightly).

I suggest that the text should be more explicit that we are talking about 
properties declared as transitive, and subproperty relationships declared 
using rdfs:subPropertyOf or owl:samePropertyAs.

[[
The explicit subproperties of a property are found by 
looking at all specified superproperties and all specifed 
EquivalentProperties, and taking the appropriate transitive 
closure.

Individual properties that are either specified as transitive or for which 
there is an explicit subproperty that is specified as transitive may not have 
cardinality conditions expressed on them, either in restrictions or by being 
functional, or inverse functional. This is needed to maintain the 
decidability of the language. 
]]

I suggest a reference to Ian's paper would be appropriate here.

Jeremy

Received on Friday, 3 January 2003 16:57:15 UTC