Lehman's laws in agile and non-agile projects

Software team leaders and managers must make decisions on what type of process model they will use for their projects. Recent work suggests the use of agile processes since they promote shorter development cycles, better collaboration, and process flexibility. Due to the many benefits of agile processes, many software organizations have shifted to using more agile process methodologies. However, there is limited research that studies how agile processes affects the evolution of a software system over time. In this paper, we perform an empirical study to better understand the effects of using agile processes. We compare two open source projects, one of which uses a tailored agile process (i.e., Xtreme Programming) and another that has no formal process methodology. In particular, we compare the two projects within the context of Lehmans Laws for continuing growth, continuing change, increasing complexity, and conservation of familiarity. Our findings show that all four of the laws held true for the project that uses an agile process and that there are noticeable differences in the evolution of the two projects, many of which can be traced back to specific practices used by the agile team.