Adding confidence intervals to the NESMA Functional Size Estimation Method

In many projects, software functional size is measured via the IFPUG (International Function Point Users Group) Function Point Analysis method. However, applying Function Point Analysis using the IFPUG process is possible only when functional user requirements are known completely and in detail. To solve this problem, several early estimation methods have been proposed and have become de facto standard processes. Among these, a prominent one is the ‘NESMA (NEtherlands Software Metrics Association) estimated’ (also known as High-level Function Point Analysis) method. The NESMA estimated method simplifies the measurement by assigning fixed weights to Base Functional Components, instead of determining the weights via the detailed analysis of data and transactions. This makes the process faster and cheaper, and applicable when some details concerning data and transactions are not yet known. The accuracy of the mentioned method has been evaluated, also via large-scale empirical studies, showing that the yielded approximate measures are sufficiently accurate for practical usage. However, a limitation of the method is that it provides a specific size estimate, while other methods can provide confidence intervals, i.e., they indicate with a given confidence level that the size to be estimated is in a range. In this paper, we aim to enhance the NESMA estimated method with the possibility of computing a confidence interval. To this end, we carry out an empirical study, using data from real-life projects. The proposed approach appears effective. We expect that the possibility of estimating that the size of an application is in a range will help project managers deal with the risks connected with inevitable estimation errors.