Experiences in the inspection process characterization techniques

Implementation of a disciplined engineering approach to software development requires the existence of an adequate supporting measurement & analysis system. Due to demands for increased efficiency and effectiveness of software processes, measurement models need to be created to characterize and describe the various processes usefully. The data derived from these models should then be analyzed quantitatively to assess the effects of new techniques and methodologies. In recent times, statistical and process thinking principles have led software organizations to appreciate the value of applying statistical process control techniques. As part of the journey towards SW-CMM® Level 5 at the Motorola Malaysia Software Center, which the center achieved in October 2001, considerable effort was spent on exploring SPC techniques to establish process control while focusing on the quantitative process management KPA of the SW-CMM®. This paper discusses the evolutionary learning experiences, results and lessons learnt by the center in establishing appropriate analysis techniques using statistical and other derivative techniques. The paper discusses the history of analysis techniques that were explored with specific focus on characterizing the inspection process. Future plans to enhance existing techniques and to broaden the scope to cover analysis of other software processes are also discussed.