Impact of schedule estimation on software project behavior
Tarek K. Abdel‐Hamid, Stuart Madnick · DSpace@MIT (Massachusetts Institute of Technology) · 2011
Schedule estimation of software development projects has historically been, and continues to be, a major difficulty associated with the management of software development.Research efforts attempting to develop "better" estimation tools need to address two important issues: first, whether a more accurate estimation tool is necessarily a better tool, and secondly, how to adequately measure the accuracy of a (new) estimation method.Our objective in this paper is to address these two issues.An extensive series of industry interviews and literature study culminated in the develop- ment of a system dynamics model of software development project management.The model served as a laboratory vehicle for conducting simulation experiments on the impact of schedule estimation on software project behavior.The study produced two interesting results: first, that a different estimate "creates" a different project.The important Implication that follows from this is that (new) software estimation tools cannot be adequately judged on the basis of how accurately they can estimate historical projects.Secondly, that a more accurate estimate is not necessarily a "better" estimate.