Investigating patterns and task type correlations in open source

This paper describes an investigation using data from a number of open source mailing lists. A methodology is described that details the procedure used to extract program summaries posted to mailing lists by experienced programmers as well as the type of analysis involved. The data are analyzed for patterns and correlations between task type and the program comprehension scheme categories used. When compared to laboratory style empirical studies, an advantage is gained allowing stronger ecological valid insights into programmer comprehension