Human and program factors affecting the maintenance of programs with deployed design patterns

ContextPractitioners may use design patterns to organize program code. Various empirical studies have investigated the effects of pattern deployment and work experience on the effectiveness and efficiency of program maintenance. However, results from these studies are not all consistent. Moreover, these studies have not considered some interesting factors, such as a maintainer’s prior exposure to the program under maintenance.ObjectiveThis paper aims at identifying what factors may contribute to the productivity of maintainers in the context of making correct software changes when they work on programs with deployed design patterns.MethodWe performed an empirical study involving 118 human subjects with three change tasks on a medium-sized program to explore the possible effects of a suite of six human and program factors on the productivity of maintainers, measured by the time taken to produce a correctly revised program in a course-based setting. The factors we studied include the deployment of design patterns and the presence of pattern-unaware solutions, as well as the maintainer’s prior exposure to design patterns, the subject program and the programming language, and prior work experience.ResultsAmong the factors under examination, we find that the deployment of design patterns, prior exposure to the program and the presence of pattern-unaware solutions are strongly correlated with the time taken to correctly complete maintenance tasks. We also report some interesting observations from the experiment.ConclusionA new factor, namely, the presence of pattern-unaware solutions, contributes to the efficient completion of maintenance tasks of programs with deployed design patterns. Moreover, we conclude from the study that neither prior exposure to design patterns nor prior exposure to the programming language is supported by sufficient evidences to be significant factors, whereas the subjects’ exposure to the program under maintenance is notably more important.

► Six human and programs factors for the effective use of design patterns are studied.
► Maintainers may not develop pattern-aware solutions even if they exist.
► The presence of pattern-unaware solutions affects the time to complete a correct change task.
► A prioritized list of the factors is proposed for further study.