> 1) The State Interface which is inside the TokenEvaluationEnging
> interface (TokenEvaluationEngine.java, SESAT query API) has two
> interesting functions/signatures: getKnownPredicates and
> getPossiblePredicates. What's the difference between these two?

Predicates are stored as meta-data within each Clause.
Each Clause is immutable and used within a flyweight pattern 
implementation. That is the one clause could be used simultaneously 
across many queries, in different positions within the queries, across 
different request.

Therefore any predicate which is dependent on position within the query 
(or dependent on any other property external to the clause) can not be 
"known" to be true. But it could be "possible".

For example MAGIC predicates only are true when they are the first word 
in the query.
But when the clause is constructed it has no relation to the query 
requesting it to be created.
The "possible" predicates must be checked, evaluated, again with the 
query also supplied. You should be able to see this pattern in various 
places.

~mck

-- 
"Physics is the universe's operating system." Steven R Garman
| www.semb.wever.org | www.sesat.no | www.sesam.no |

_______________________________________________
Kernel-development mailing list
[email protected]
http://sesat.no/mailman/listinfo/kernel-development

Reply via email to