Greetings! I read through the manual, but it is a little too terse regarding when one might choose Stateless versus Stateful sessions. My initial testing has been with Stateful sessions and these work nicely but memory use explodes as the number of facts added increases much beyond 100. So, my questions: 1. Do stateless sessions execute faster than stateful? I'm assuming that even if they don't, they use memory more effciently? 2. I read that stateless sessions do not allow for facts to be modified yet I saw an example online which was using "insert()" in the DRL file. Will this work if there is a rule that has not already fired but might after the insert? In a stateless session will a rule only fire once and ignore new facts? (I believe so). 3. It also seems that a stateless session has no agenda and therefore will simply fire the first rule that matches so that salience must be used to prioritize rules. Is this correct? 3. The difference between a stateless session and a sequential stateless session is also unclear (to me) from the manual. Can someone explain it? What I am looking to build is essentially a filter, something that gets a bunch of input facts and applies the rules to see if any alarms should be raised based on the inputs. My test programs have been using a global which tracks the alarms raised. I just look at it after the engine has run and use it as my output. Each run of the rules engine is separate from the previous run, ie, there is no accumulation of "knowledge", which is why I think a stateless session is the way to go but I'm not sure. Any help appreciated! Ron _________________________________________________________________ The i’m Talkathon starts 6/24/08.  For now, give amongst yourselves. http://www.imtalkathon.com?source=TXT_EML_WLH_LearnMore_GiveAmongst