[rules-users] Request user feedback during rules execution

[Edson Tirelli]

   Katerina,

   I will give you my personal opinion, and it is important that you 
understand that it is not based in any literature or others experience. 
Just my own professional experience.

   While it is possible to do that in the RHS of a rule, I would not do 
it. The main reason for that is that it would make everything more 
difficult to develop/test/maintain. You would increase the coupling 
between your system's components and decrease their flexibility.
   When designing my own systems, I tend to specialize components with 
well defined roles. When working with rules, I usually let rules take 
the decisions, but let other specific component react to those decisions 
(in this case, manage the flow of the application). I think about the 
rules engine as a "service" in a SOA architecture (if you prefer to 
think like that), in a way that if more information from the user is 
needed, what the rules will do is to answer to the service call "saying: 
'more information needed". The real controller component (the one that 
manages the application flow) will understand that message and ask the 
user for more info. Once the user adds the info, the controller asserts 
new info into the working memory and fire rules again (or call the 
service again, if you prefer).

   The advantages for me in such approach is that a small application 
can use such "rules service" with a house made controller and do fine 
with it, but if you need to leverage that controller for something more 
complex like managing a workflow or business process, you can replace it 
for a BPM engine (like jBPM) and your rules will never notice (better 
decoupling). Also it becomes really easy to implement unit tests over 
your rules, while if you increase coupling with GUI components, you 
would need to start creating mocks for testing your rules.

   So, my answer is model your rules component as a passive service and 
have a controller to be the active flow manager.

   Hope it helps.

   []s
   Edson

kchantzara at costamare.com wrote:

>Hello,
> 
>I am thinking about using JBoss Rules as the mid-tier of a web application.The aim of this system is to conclude on some results based on the rules and on some data provided by the user by the web interface. The problem is that apart from the initial input, some more user information might be needed during the rules workflow. The user response might result in asserting new objects into the working memory, firing new rules etc. My question is, would it be possible to request user information as part of the RHS of a rule and 'halt' the firing of rules until this information is provided? 
> 
>Thanks in advance,
>Katerina
>
>_______________________________________________
>rules-users mailing list
>rules-users at lists.jboss.org
>https://lists.jboss.org/mailman/listinfo/rules-users
>
>  
>


-- 
 Edson Tirelli
 Software Engineer - JBoss Rules Core Developer
 Office: +55 11 3124-6000
 Mobile: +55 11 9218-4151
 JBoss, a division of Red Hat @ www.jboss.com