I'm new to this list, but I have a question about the performance costs of
removing a rule. I know that if I have, for example, 100 rules that there is
an API to remove just one of them. However, what I don't know is how
efficient it is to do this. Does removing just one rule clear out all of the
existing inferences and then recompute the working memory from scratch
(perhaps by recompiling the remaining 99 rules)? Or, alternatively, does
Drools have some efficient way of just removing that one rule along with
only the specific items that were inferred using it? Does anyone know?
View this message in context: http://www.nabble.com/Performance-on-Rule-Removal-tp21240363p21240363.html
Sent from the drools - user mailing list archive at Nabble.com.
Can anyone point me to some solid examples for mixing processes with
human interaction? I've been trawling through the codebase and it
looks like drools is geared to support all kinds of different process
mappings (jpdl, ruleflow, bpel, etc.) on the core engine. Looks very
cool and potentially powerful, but I'm trying to get my head around
how to use it with non-dev users.
It seems the design philosophy is to model the domain as Java objects
then define rules and flows that use these objects. Trouble is I
can't see the interface points to create a UI (rich client or
whatever) to have a human user take required actions for various
processes. This could be as simple as 'get task and data blob', do
something to it, and post it back (correlating the process ID).
For point of reference, I was quite familiar with how Weblogic
Integration managed processes and human tasks, and also have checked
into Singleshot (nice idea, but looks incomplete). At the moment
using something like Bonita appears to be the best bet, but with some
docs/examples/configuration drools could fit the bill.
Then of course I'm hoping there's the infrastructure in place to
activate/passivate the processes (memory to disk) on requests/
responses from the human users
thanks in advance,