A phased approach to the PSP

Author(s): Jeffrey S. Holmes and Bonnie E. Melhart
Venue: Sixteenth Annual Pacific Northwest Software Quality Conference
Date: October 13 - 14, 1998

Sample Size: 1
Class/Experience Level: Professional
Data Collection Method: Project Artifact(s)


Link: http://www.pnsqc.org/proceedings/pnsqc98.pdf
Type of Experiment: Experience Report

This paper proposes a method by which PSP can be introduced into industry. The paper then goes on to describe the results from an attempt to use this approach. The industrial setting was Motorola, and the items being coded were real-time embedded applications. The method described is a phased approach where pieces off the process are introduced one at a time rather than using the “big-bang” approach.
One interesting finding was that even in the cases where the time estimates (man hours) were solid, other organizational factors kept the estimated calendar date from being hit. One such organizational factor was changing priorities. This certainly speaks to the difference in PSP studies performed in academic settings (where there are no interruptions due to real world factors) and industrial settings.
The data results that were displayed simply compared the estimated size, time, and defects with the actual. Discrepancies between the estimates and actual were discussed and rationalized. This isn’t a strong empirical study as there is little more than this discussion of collected values.