Software Size Estimation of Individual Projects

Cuauhtémoc López-Martín, Arturo Chavoya, and María E. Meda-Campaña

Keywords

Software development, software size estimation, programming languages, expert judgment

Abstract

Background: Software project size is often used as independent variable for predicting dependent variables such as effort, schedule, costs or risks of software projects. The better the size estimation accuracy, the better the prediction accuracy of dependent variables. The size of each project is usually estimated in terms of the number of lines of code for the programming language in which the project will be coded. This estimation is typically done using expert judgment techniques or applying prediction models. Hypothesis: There is a statistically significant difference amongst size prediction accuracy of projects by the object-oriented programming language used, when they are estimated by expert judgment from specified requirements in natural language. Method: A population of 1,414 individual software projects was developed by 202 practitioners. Each project had its own specified requirements in natural language, and each one was developed within a controlled experiment and following a disciplined process. A sample of 676 projects developed in C++ or Java was selected for this study. Results: There was a statistically significant difference between size estimation accuracy for C++ and Java at a 95% level of confidence. Conclusions: The size estimation accuracy of software projects coded in Java was better than the estimation accuracy of projects coded in C++.

Important Links:



Go Back