Continuous Software Process Improvement through Statistical Process Control

Measurement based software process improvement is nowadays a mandatory activity. This implies continuous process monitoring in order to predict its behaviour, highlight its performance variations and, if necessary, quickly react to it. Process variations are due to common causes or assignable ones. The former are part of the process itself while the latter are due to exceptional events that result in an unstable process behaviour and thus in less predictability. Statistical Process Control (SPC) is a statistical based approach able to determine whether a process is stable or not by discriminating between the presence of common cause variation and assignable cause variation. It is a well-established technique, which has shown to be effective in manufacturing processes but not yet in software process contexts. Here experience in using SPC is not mature yet. Therefore a clear understanding of the SPC outcomes still lacks. Although many authors have used it in software, they have often not considered the primary differences between manufacturing and software process characteristics. Due to such differences SPC cannot be adopted "as is" but it must be tailored. In this sense, I propose an SPC-based approach that reinterprets SPC, and applies it from a Software Process point of view.