An evaluation of functional size measurement methods

Abstract

Background: Software size is one of the key factors that has the potential to affect the effort of software projects. Providing accurate software size estimation is a complex task. A number of functional size measurement (FSM) methods have been proposed to quantify the size of software based on functional user requirements (user perspective). Function point analysis (FPA) was the first proposal for a FSM method and it is one of the most accepted FSM methods in the industry. Automated Function Point (AFP) method state the guidelines for automating FPA counting from software source code. Objectives: This paper reports on an experiment that compares FPA and AFP. The goal is to evaluate the measurement process on a range of performance and adoption properties such as accuracy, reproducibility, efficiency, perceived easy to use, usefulness, and intention to use. Methods: A controlled experiment was conducted to compare the two methods. Statistical analyses were conducted to find differences between the methods regarding performance and adoption properties. Results: The functional size results between the FPA and AFP methods were similar (MMRE 6-8%). Productivity rate was about the same reported for the industry (43.4 FPA/h, 37.8 AFP/h). There were no significant differences between the methods for functional size estimation, reproducibility, and accuracy. Limitations: This is an initial experiment of a work in progress. The limited sample size and nature of the subjects may influence the results. Conclusions: These results support the claim that AFP produces similar measurement results that FPA. The automation of the AFP method could produce more consistent measurement results in conformance with the FPA counting guidelines. An automated and quick FSM counting method will increase the adoption of this metric in industry. Further research is needed to conclude more on some perceived adoption properties.