Thank you Michael. By the way thank you for Drools 3 it is looking really good. Oscar
On 4/9/06, Michael Neale <[EMAIL PROTECTED]> wrote: > > no that sounds fine for your application. I assume you have a cached > single > copy of RuleBase (which is where most of the "warm up" time would go). You > only need to share working memories if you are keeping state in them that > is > needed between transactions (as otherwise, you are basically asserting the > same objects over and over). > > On 4/8/06, Luis Trigueiros <[EMAIL PROTECTED]> wrote: > > > > Dear Drools community, > > > > I am currently working on a project using Drools 2.1 as the rule engine > > provider. > > > > It is a web-based application that uses Drools to implement dynamically > > configured logic controlling what the user is allowed to do with domain > > objects. > > > > As such Drools is controlling: forms validation, users ability to access > > the > > object and his fields, and loads of other stuff it is all very much all > > rules driven. > > > > At the moment during a web request handling cycle we do not reuse the > > working memory. In that the same working memory is created multiple > times > > motivated by: > > > > · Thread-safety concerns > > > > · Rules are crosscutting in the system and invoked by several > > components that are unaware of each other. > > > > Given the previous facts I would like to posse following questions: > > > > · Are we making improper use of Drools by using it this way? > > > > o Does Drools remember the evaluation previous object state? > > > > o And if the working memory associated with an object were to be reused > > how > > could Drools be notified of the change (at the moment were are asserting > > the > > all object to working memory and > > > > not using application data) > > > > I understand that these questions are may sound a bit vague. But maybe > > some > > of you have already face then while integrating a rule engine in a web > > application. > > > > > > > > Best regards, Oscar > > > > > >
