Abstract
Measurements calculated by dividing project size by development effort are commonly used to evaluate productivity. Although such measurements are simple and clear to understand, the overall accuracy is very poor. In order to address this poor accuracy, we have introduced four methods for productivity evaluation based on combinations of source lines of code and development effort. We have completed a quantitative review of the methods in terms of accuracy and clarity using actual data from completed projects and studied possible implementation in various occasions. In this paper, we describe the results of the review and study.
