Conclusion


In this chapter, we have identified two process models for conducting micro-incremental testing (that is, incremental testing with short iteration cycles). The first model does not dictate an ordering for constructing methods or for developing and running test cases but is specific about the structure of test cases and the steps required at each iteration. This model could be viewed as a formalization of the process described in [Beck+1998].

The second model requires the identification of multiple test cases at each iteration of the development cycle and then provides guidance as to the order in which those test cases are tested and their methods implemented. With this model, seven test cases could be run as the first five methods are developed; in contrast, only two test cases could be run as the first five methods are developed with the ordering that we chose under the ad hoc model. Our heuristic simply and efficiently gives an ordering over the set of possible methods that is consistent with the XP practice "Do the simplest thing that could possibly work." Our heuristic seems to hold particular promise in the context of certain refactorings in which the extracted methods, attributes, and test cases are known before the refactoring is initiated. Future work will consider the impact of discovered tests (where the need for a test becomes apparent while code is being developed), testing partially completed methods, and application of this technique to testing during refactoring.



Extreme Programming Perspectives
Extreme Programming Perspectives
ISBN: 0201770059
EAN: 2147483647
Year: 2005
Pages: 445

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net