in a machine learning situation where the rule engine is running for hours,
days or weeks, it is possible the "honest politician" fact could be asserted
and retracted. It's perfectly acceptable for the rule to fire more than once
if the state changed such that at time T

T1 - 1 honest politician
T2 - 0 honest politician
T3 - 2 honest politician
T4 - 0 honest politician
T5 - 3 honest politician
T6 - 2 honest politician
T7 - 0 honest politician

The business cases I've seen do not work this way. hope that helps

peter

On 3/24/06, Peter Van Weert <[EMAIL PROTECTED]> wrote:
>
>
> > rule "politician"
> >   when
> >     not exist politician (honest == true)
> >   then
> >     incarcerate all politicians
> >
> > In the not join case, there may be multiple matches that will fire. in
> the
> > Not exist case, it should only fire that rule once.
> >
> Ok, and what about the following scenario:
> - at some point there is no honest politician (==> rule fires)
> - next a honest politician is asserted gets asserted (==> rule is no
> longer applicable)
> - then this politician is removed again (or becomes dishonest)
>
> What would/should happen in this case? Would/should the rule fire once
> more or not?
>
>

Reply via email to