In this chapter we have covered different ways to represent our knowledge in Drools.

For a tailored language for specific use cases, DSL/DSLR could be a solution that hides the complexity of the DRL syntax behind a language that is more familiar to an SME.

For situations where a big set of rules with a common structure is required, decision tables could be a perfect fit. The big advantage of decision tables is that they are very concise and user-friendly.

When more flexibility in the structure of the rules is required, rule templates could be a good help. The out of the box support for data sources such as spreadsheets, objects, or even a SQL store makes rule templates a very interesting option.

If we are dealing with non-rule based knowledge such as neural networks, decision trees, or scorecards, we must consider the use of the drools-pmml module. Maybe this is not the best performance solution but it allows us to easily integrate these models with our current rule-based solution.

Now...