An empirical study on the factors affecting software development productivity

Background - Software development productivity is widely investigated in the Software Engineering literature. However, continuously updated evidence on productivity is constantly needed, due to the rapid evolution of software development techniques and methods, and also the regular improvement in the use of the existing ones. Objectives -The main goal of this paper is to investigate which factors affect productivity. It was also investigated whether economies or diseconomies of scale exist and whether they may be influenced by productivity factors. Method -An empirical investigation was carried out using a dataset available at the software project repository ISBSG. The major focus was on factors that may affect productivity from a functional point of view. The the conducted analysis was compared with the productivity data provided by Capers Jones in 1996 and 2013 and with an investigation on open-source software by Delorey et al. Results -This empirical study led to the discovery of interesting models that show how the different factors do (or do not) affect productivity. It was also found out that some factors appear to allow for economies of scale, while others appear to cause diseconomies of scale. Conclusions -This paper provides some more evidence about how four factors, i.e., programming languages, business areas, architectural types, and the usage of CASE tools, influence productivity and highlights some interesting divergences in comparison with the results reported by Capers Jones and Delorey et al.