[rules-users] data manipulation

Hi all, We recently implemented Drools in a mortgage application process. The primary task of the business rule engine was to determine the fraud risk of an application. The application is delivered in XML format. We noticed that we were implementing a lot of data manipulation actions in the drl file itself. For example, the calculation of total income by calculating the sum of a number of elements or calculating the distance between two geo-locations based on zip codes. The actual evaluation rules acted on this 'derived' information. When looking back the drl file contained a lot of code and looked pretty complicated compared to the actual evaluation rules. My question is, what is your opinion on data manipulation or transformation within the drl file? If the drl file is not the best place for this, what would you suggest for data manipulation? Thanks, Betto