Dynamic rich-data capture and analysis of debugging processes

This paper proposes a methodology for the study of program comprehension and debugging through the capture and analysis of rich process data. A software debugging environment with enhanced functionality is used to capture these data and a mixture of qualititative and quantitative approaches is employed to analyze them. The functionality added to the software debugging environment allows it to record the programmers' verbalisations, their focus of visual attention and their keyboard and mouse actions. These synchronous data are analysed to build a model that relates debugging expertise to strategy in terms of representation coordination and individual differences in representation use.