14. Issues with Rule Testing and Performance

Target audience Application architect, software architect, developer, business analyst, nontechnical audiences can skip Sect. 14.6 In this chapter you will learn The rule testing approach and how it fits into the different testing phases used in traditional software development How to use a test-driven approach to implement business rules How using SCA can help build the application and a test framework by iterations How performance could be impacted by business rules application, and what to look for during design and testing activities Testing does not end once in production, especially with rule-based applications What is rule semantic consistency checking, and what the different search patterns are The problem of tracing and logging Key points Adopting a test-driven development helps to implement the rules by first looking at their intent. Testing business rules is about business, so the test has to be designed to prove the impact to the business. The test framework should be light and not too rigid on the assertion of the expected results, as some results are computation driven. Performance tests should measure all the potential bottlenecks, like the messaging layer, the data conversion, the data access, and the rule execution. Rule engines are optimized to run fast, and sometimes data access is the issue. As in database, join operations are costly. Testing activities remain important during the maintenance of the application. Inconsistencies in a ruleset may happen over time, but they have to be corrected as soon as possible. Some BRMS offer consistency checking features to search for rules never selected, rules that make other rules redundant, rules with equivalent conditions …