Refactoring Needs and Extreme Programming

In this study, we used a near-clone detector to search for refactoring opportunities in two implementations of the same software. The first was developed in a typical manner. The second was developed subsequently by members of the same research team employing several extreme programming (XP) practices, including test-driven development and pair programming. We searched the two resulting Java packages and the test code for the second project for pairs of methods with large alignment of their parse trees. We examined these pairs to determine whether the pair was truly refactorable, and the type of refactoring needed. Initial results indicate that refactoring opportunities persist in all three code bases, but they differ in kind and seriousness. XP and other “ agile” methodologies are growing in number of practitioners and may become a significant factor in industrial software development over the next decade. As such, our study may signal to refactoring specialists a coming shift in the types of refactoring tools that will be needed by XP teams in the future, and to XP practitioners the nature of continued opportunities for refactoring that many XP developers are missing.