Inoculating Novice Software Designers with Expert Design Strategies
David Wright · 2020
Abstract Inoculating Novice Software Designers with Expert Design Strategies Abstract Computer science and software engineering students need to develop the knowledge andskills necessary to build large, complex, and reliable software systems to solve problems thatare ill-structured, incompletely defined, and full of apparently conflicting information. Theapproaches employed by novices to solve these types of problems are significantly differentthan those used by expert designers. The Principles, Patterns, and Process Framework(P 3 F ) is proposed as a solution to the problem of encapsulating expert design strategiesand making them accessible to novice designers. In particular, the P 3 F is designed to helpnovices build and maintain a systemic view of a design problem, efficiently navigate andmanipulate the problem/solution space, and make effective design decisions. This paper presents the results of an interrupted time-series experiment conducted tostudy the effects of introducing the P 3 F to novice software designers. Data collectedincluded subjects’ self-assessments of their design behaviors, verbal protocols of the subjectsworking on complex software system design problems, and the artifacts representing theirsolutions to these problems. The self-assessments were compared against the behaviorsobserved in the recorded protocols to characterize the subjects’ design behavior before andafter the introduction of the P 3 F and identify changes attributable to their use of theFramework. The solution artifacts were evaluated against a design quality rubric and theseresults were analyzed to identify improvements in the quality of the solutions that correlatedwith the introduction of the P 3 F . The results of this study suggest that the Principles, Patterns, and Process Frameworkcan help novice software designers adopt expert design strategies and improve their personaldesign skills. Two of the five subjects that completed the study demonstrated significanttransformations in their construction and navigation of their view of the design problem andevolving solution as well as an improvement in the quality of the solutions they produced.Another two subjects adopted, to a lesser extent, aspects of the expert strategies incorporatedin the P 3 F while retaining much of their novice behaviors. The fifth subject showed littleor no behavioral change or difference in solution artifact quality. While there is a small but significant body of research studying expert and novice designbehaviors, there have been few attempts to bridge the gap to cultivate expert-like behaviors innovice software designers. This work contributes such an attempt combined with an effort tomeasure the change induced by the introduction of the P 3 F . Furthermore, because the designof the Framework drew from a broad range of design-intensive disciplines, the P 3 F can beadapted to suit a variety of design environments. The positive results obtained in this studyinvite further investigation and refinement of the P 3 F in the realm of engineering design.