Re: Categorization of Predicates

"Neo555" <member_at_dbforums.com> wrote in message news:2005715.1036383995_at_dbforums.com...
>
> I am developing a database that attempts to processes a new predicate
> introduced into the db in a manner similar to it's category. Has someone
> already categorized predicates before? I want to categorize predicates
> that combine two things (ie thing1 predicate thing2) from a very
> general and logically processable point of view. Are the below
> categorizations correct?
>
> BIDIRECTIONAL PREDICATES:
> t1 and t2 can be exchanged and the proposition remains true.
> For example:
>
> John agreesWith Bob.
> Bob agreesWith John.
>
> Mary isNextTo John.
> John isNextTo Mary.

And a relvar holding such propositions must have the constraint

    IF (t1, t2) THEN (t2, t1)
defined

> UNIDIRECTIONAL PREDICATES:
> Exchanging t1 and t2 results in a false proposition.
> For example:
>
> John ate the apple. Its negation, The apple ate John, is false.
> Mary cut the wood. Its negation, The wood cut Mary, is false.

And a relvar holding such propositions must have the constraint

    IF (t1, t2) THEN NOT (t2, t1)
defined

> HIERARCHAL/TRANSITIVE PREDICATES:
> If we have: t1 predicate t2.
> And we have: t2 predicate t3.
> Then: t1 predicate t3, is also true.
> For example:
>
> John isA Person. Person isA Thing. John isA Thing.
> Bread isPartOf sandwich. Sandwich isPartOf lunch. Bread isPartOf lunch.

And a relvar holding such propositions must have the constraint

    IF (t1, t2) AND (t2, t3) THEN (t1, t3) defined

So this is more of a categorisation of constraints as of predicates per se?

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Mon Nov 04 2002 - 21:18:13 CET