Waiting in software projects: an exploratory study

Abstract

In preparation for a larger investigation, Bradac, Perry and Votta conducted a time diary study of how a developer spent their time over a 30-month period. They conducted their study to better understand, and ultimately try to reduce, project duration. We re-analysed some of the evidence presented in Bradac et al. s paper, and this led to a modification to one of Bradac et al. s conjectures viz. that, at the level of an individual developer, waiting is more prevalent--during the beginning of the project. Using two projects at IBM Hursley Park, we then investigated whether the revised conjecture applied at a higher level of the project i.e. at the level of process areas. Our finding of behaviour at the level of process areas is inconsistent with our revised version of Bradac et al. s conjecture i.e. we found that waiting is more prevalent during the end of the project. Further analysis confirmed similar behaviour for reports of poor progress and outstanding work. Other evidence from the two projects lead us to believe that project urgency affects the reporting made, by managers, of the status of process areas. We relate our findings to the Deadline Effect, and discuss our findings with regards to effort.