[drools-research] Detection semantics

[Oroess]

Hello everyone.

I'm currently evaluating Drools as a part of my bachelor thesis. And I'm
having trouble to categorize Drools regarding detection semantics of complex
events.

I understand that there are two possibilities of timestamping events: just
time points or time intervals (where a duration of 0 represents a time point
event).
Now, Drools definitely implements time intervals, no doubt there.

Then I understand that there are semantic differences in detecting complex
events. This problem stems from composition-operator based languages. If a
complex event (represented by its pattern) is detected, it could get
assigned the occurrence time as timestamp of the last event which was needed
for the pattern OR it could get assigned an interval of the occurrence times
of the first and the last event needed for the pattern.

I have read a lot of papers which describe this problem. I have read the
papers specified in the Fusion docs and even some more like

1. Walzer K, Groch M, Breddin T. Time to the Rescue - Supporting Temporal
Reasoning in the Rete Algorithm for Complex Event Processing. 2008:635-642.

and

1. Walzer K, Breddin T, Groch M. Relative temporal constraints in the Rete
algorithm for complex event detection. Proceedings of the second
international conference on Distributed event-based systems - DEBS ’08.
2008:147. Available at:
http://portal.acm.org/citation.cfm?doid=1385989.1386008.

But I can't seem to find a statement how Drools ultimately manages this.

In Drools, there seems to be no such problem at all. Rules (LHS) can also
feature facts which have no notion of time, and the user is able to set the
timestamp (and duration) manually when inserting a new event through the RHS
of a rule. Since its a rule based language, and the RETE algorithm works
differently than the implementations of composition-operator based
languages, I infer that there is no problem with detection semantics.
Am I right here?

Thanks in advance
Kind regards
Oliver


-- 
View this message in context: http://drools-java-rules-engine.46999.n3.nabble.com/Detection-semantics-tp2365940p2365940.html
Sent from the Drools - Research mailing list archive at Nabble.com.