search

own-pt / cl-krr

Environment for knowledge representation, reasoning, and engineering.
Apache License 2.0
4 stars 3 forks source link

Redundant relativization #1

Closed fcbr closed 5 years ago

fcbr commented 8 years ago

This formula:

(=>
   (exhaustiveDecomposition ?CLASS ?ROW0 ?ROW1)
   (and (instance ?CLASS Class) 
        (instance ?ROW0 Class) 
        (instance ?ROW1 Class)))

was relativized as:

(forall (?CLASS ?ROW1 ?ROW0)
 (=>
  (and (instance ?ROW0 Entity) (instance ?ROW0 Class)
       (instance ?ROW1 Entity) (instance ?ROW1 Class)
       (instance ?CLASS Class) (instance ?CLASS Entity))
  (=> (exhaustiveDecomposition3 ?CLASS ?ROW0 ?ROW1)
   (and (instance ?CLASS Class)
        (and (instance ?ROW0 Class) (instance ?ROW1 Class))))))

Which is clearly wrong: the purpose of the formula is to instantiate ?CLASS, ?ROW0, and ?ROW1 as instances of Class, so they can't be added as restrictions in the antecedent.

arademaker commented 8 years ago

Can we generalize it so all (instance ...) in the conclusion of the rules should not contribute to types restrictions ?

fcbr commented 8 years ago

Yes, that's the idea.

arademaker commented 5 years ago

Close in favor of #24. Same issue.