Teaching and Using PSP in a Software Engineering course An Experience Report

Author(s): K Venkatasubramanian, S Bastin Tony Roy, Muralikanth V Dasari
Venue: Software Engineering Education and Training Annual Conference
Date: 2001

Sample Size: 36
Class/Experience Level: Graduate Student
Participant Selection: graduate student classwork
Data Collection Method: Project Artifact(s)


Link: http://waitaki.otago.ac.nz/~tonyr/papers/Teaching_and_using_PSP_in_a_sof...
Type of Experiment: Experience Report

This is yet another PSP experience report based on the 10 PSP programs. The reported experience is much like others that are held in an academic setting and use the standard 10 PSP course programs to collect data against. In this study it was observed that the total number of defects found during the compile and test phase of development fell as the students progressed through the 10 programs. This is a goal of the PSP process because there’s a focus on finding defects early in the software development cycle. Defect removal rate increased throughout the class; particularly after PSP program 7 was written. Another result that was measured by this study is that overall programmer productivity didn’t drop off. This is significant because the PSP does add overhead to the programming process in the form of data collection and analysis. The PSP concept is that your productivity shouldn’t drop off, and that the additional analysis will make you a better programmer over the long haul. Finally the students’ ability to estimate effort and program size became better throughout the study.

The authors suggest that the main problems faced in getting students to use PSP permanently are the discipline that it takes, and the lack of tools that support the PSP process.