The Personal Software Process in Practice - Experience in Two Cases over Five Years

Author(s): Georg Grutter and Stefan Ferber
Venue: 7th European Conference on Software Quality, ECSQ
Date: 2002

Sample Size: 2
Class/Experience Level: Professional
Participant Selection: self-selected
Data Collection Method: Project Artifact(s)


Type of Experiment: Experience Report

In this study the authors relate their personal experience with the PSP in a professional setting. The authors were the subjects of the study. Both authors were graduates of the PSP course. The experiences reported are those of both authors while working on separate projects. Interestingly both projects had teams of up to 10 engineers, but the two authors were the only ones practicing the PSP.

In one of the projects, the author found that estimating the number of defects that would be injected into the program turned out to be very useful in tuning the intensity of code reviews. Also in this case the time and size estimating was very successful. In the other case the size and time estimating never stabilized, but the authors hypothesize that this was a result of a change in the nature of the programming being performed. These changes reduced the value of the data that had been collected for past phases. The change in the nature of the programming was in response to the growing size off the project, and a need for test code to be generated to automate testing of the application. The two cases show varied results, but the authors justify the variation by explaining events that’s impacted the overall project. This was an interesting paper because it occurred in an industrial environment rather than an academic setting. Many published PSP studies simply use data from the PSP course. The problem with this academic approach is that researchers rarely get to see how well the PSP can adapt to the more dynamic, uncontrolled environment of industry. In this paper there is a hint of how the PSP performs in a “live” situation.