The vital few and trivial many: an empirical analysis of the Pareto distribution of defects

Abstract

The Pareto Principle is a universal principle of the “vital few and trivial many”. According to this principle, the 80/20 rule has been formulated with the following meaning: For many phenomena, 80% of the consequences originate from 20% of the causes. In this paper, we applied the Pareto Principle to software testing and analysed 9 open source projects (OSPs) across several releases. The results show that a small number of files account for the majority of defects, even across several releases. In contrast, there is no evidence that this small part of files containing most of the defects also makes up only a small part of the system's code size. While this is not the first study about the Pareto Principle, it adds to the body of empirical body of knowledge wrt. software defects.