Another noob question:
Is there a generally understood way of implementing the
concept of a time interval rule?
That is, lets say I have a Drools-enabled client which is
constantly receiving information from outside resources. Most of this data (and
the consequences of its changes) is time sensitive, so I'm calling the
fireAllRules() method in a loop every minute.
And now, let's say I have a rule that in addition to being
driven by incoming data changes has to be linked to a repeating clock interval
or synched with wall time such that the rule only activates and fires if the
data meets certain parameters AND it's been exactly 1 hour since the last check
or that it will fire at the top of every wall time hour (1 PM, 2PM, 3 PM,
etc.).
Furthermore, timing may change. A successful rule activation
and firing might do something like "now change this timing so that for the next
24 hours, this rule must be examined every 30 minutes instead of every hour".
My initial reading of the Drools documetation I came across
the Duration, and Date-* keywords. Their descriptions didn't seem to fit the
bill.
I ask this because our old rules engine software that has been
retired to obsolescence allowed this kind of thing to be easily set up. Coming
at it in the Drools world seems a lot more challenging.
Thoughts? Examples?
-Allen