I see several authors have now joined the decision table debate since my article. The concept frequently mentioned here is my supposed coupling between the algorithm and the table representation. I guess for the particular article, the algorithm itself used by a vendor was not my focus. I was more trying to illustrate various useful ways of representing business rules and cared less about what the actual algorithm did. From a business user’s perspective the simplicity of representation and organization is really the important part. I am glad however this thread has been picked up and dicussed to, in the end, benefit our audience.

In some cases, one can use LEAPS algorithm if the rules fit nicely into a decision table structure. JRules has sequential evaluation, which is basically LEAPS. The real benefit of decision tables in my mind is lower learning curve for users. Whether the use case fits decision table is a different story.

Drool3 has some support for spreadsheet and it compiles the rules to RETE, so it’s straight forward.

my bias 2 cents.