Evaluating the Accuracy of Estimates: the Case of Model-based COSMIC Functional Size Estimation

Functional Size Measurement is widely used, especially to quantify the size of applications in the early stages of development, when effort estimates are needed. However, the measurement process is often too long or too expensive, or it requires more knowledge than available when development effort estimates are due. To overcome these problems, early size estimation methods have been proposed, to get approximate estimates of functional measures. In general, early estimation methods adopt measurement processes that are simplified with respect to the standard process, in that one or more phases are skipped. So, the idea is that you get –at a fraction of the cost and time required for standard measurement– size estimates affected by some estimation error, instead of accurate measures performed following the longer and more expensive standard measurement process. In this paper, we consider some methods that have been proposed for estimating the COSMIC
(Common Software Measurement International Consortium) size of software during the modeling stage. We apply the most recent methodologies for estimation accuracy, to evaluate whether early model-based estimation is accurate enough for practical usage.
The contribution of the paper is twofold: on the one hand we provide a reliable evaluation of the accuracy that can be obtained when estimating the functional size of software applications based
on UML models; on the other hand, we get indications concerning the effectiveness and expressiveness of recently proposed accuracy estimation methods.