Yuri,

   Not is a simple and cheap CE to use. In a simple analysis, the cost of
NOT CE is even lower than a JOIN, since it will try joins, but will
propagate a single tuple. So, go ahead, your second approach is the best way
to go.

   []s
   Edson


2007/7/23, Yuri de Wit <[EMAIL PROTECTED]>:

I am working on a drools application with few rules and large number
of facts. In my first design I tried to avoid excessive joins thinking
I was helping improve performance but didnt realized that I was
actually shooting myself in the foot. I was basically creating a
single facade-fact that would contain two or three diff concerns
joined under the same interface. The problem I am seeing is that for
simple things like changing the status of one of many facts would
cause that fact to be reevaluated against all the other facts.

I then realized that thinking relationally about the problem would not
only simplify my solution but also probably make a lot faster.
However, in this new and relational solution I will need to make use
of many "not" CE.

My question is: is there any cost in using "not"s that I should be
awae of? Any other words of wisdom re: improving the performance in
small rules x many facts?

thanks,
-- yuri
_______________________________________________
rules-users mailing list
[email protected]
https://lists.jboss.org/mailman/listinfo/rules-users




--
 Edson Tirelli
 Software Engineer - JBoss Rules Core Developer
 Office: +55 11 3529-6000
 Mobile: +55 11 9287-5646
 JBoss, a division of Red Hat @ www.jboss.com
_______________________________________________
rules-users mailing list
[email protected]
https://lists.jboss.org/mailman/listinfo/rules-users

Reply via email to