Empirically Guided Software Guesstimation

The state of the art in software project estimation does not deserve this reputation, but the fact remains that modern methods are time-consuming and complex. Estimation methods such as COCOMO II [3] are oriented toward the needs of large and very large software engineering projects, and may require time-consuming calibration to fit the peculiarities of an organization, development team, and application domain. On the other hand, methods such as the Personal Software Process [8] have demonstrated good results in improving small/individual project estimation accuracy, but advocates promote the need for specialized training and adherence to a highly constrained approach to development. Although these approaches have demonstrated success in organizations with the resources necessary to adopt them, the process overhead involved in