Students' perceptions of software risks

Abstract

Risk-the possibility of an unsatisfactory outcome-is an essential vehicle for a software development project to progress. Iterative and incremental process models like spiral advocate the continuous identification of the items likely to compromise the project's success and the early resolution of those top-ranked risk items. Although the concepts and principles, such as risk exposure and project top-10 risk-item monitoring, are commonly taught in undergraduate software engineering courses, little is known about how students, especially those working in agile software teams, perceive, prioritize, and try to mitigate their risks over multiple development cycles. In this paper, we report the data collected and analyzed in two semesters of a junior-level software engineering course where undergraduate students were working in agile teams to deliver 4 major working increments per semester (62 students developed Eclipse plugins in one semester and 103 students developed Android apps in the other). In both semesters, we found that not only were our students' perceived top-ranked risks remarkably different from what were previously published (including the industry-surveyed checklists in the 1990s and 2000s, as well as the ones collected from a graduate-level course), but the risk management strategies adopted by our students were inherently collaborative. We leveraged this collaborative nature to design and execute the instructor interventions. The results comparing the top risks between the two semesters show the effectiveness of the instructor interventions and suggest ways to further improve risk management in students' agile software development teams.