Extreme makeover: Bending the rules to reduce risk rewriting complex systems

Abstract

We describe our experience using XP to reimplement sophisticated, high-performance imaging software in a research environment. We focus especially on practices we used to derive value from the existing software, notably reimplementation by ransacking and conversion as learning. Our experience suggests that some of the classic 12 practices which define XP should be adjusted when there is a existing, well-structured system to serve as a guide.