The work life of software developers has often been studied with a goal to understand ways to enhance productivity. The authors extend past efforts by gathering detailed work data from 20 experienced developers via a mix of unobtrusively monitored computer interaction data (for example, applications used, keystrokes, mouse clicks) and periodic self-reports (for example, work activities, perceived productivity). The developers are volunteers from four different organizations across three countries doing routine software development work. On average, each developer reported 11 full workdays of activities.

The research study is methodologically impressive, and the authors provide several rigorous analyses that offer interesting insights. The study results find that software developers experience a highly fragmented workday with a multitude of unplanned interruptions that hamper productivity. The data show that how individual developers perceive their own productivity and what influences greater productivity varies widely, creating a complex picture of how developers work. One interesting observation is that developers can be grouped into three categories based on the times of day they are most productive: morning, afternoon, and low-at-lunch. The authors conclude with proposals for designing work processes and tool support for improving developer productivity. Overall, the research is well documented with clear descriptions of its goals, methodology, results, implications, and future research directions.