An Empirical Comparison of Java and C# Programs in Following Naming Conventions

Abstract

An important indicator of source code quality is compliance with naming conventions. It is believed that such practices improve program comprehension, which directly affects maintainability and reusability. In this paper, the authors conduct an experiment to determine how well Java and C# programs follow a set of well-publicized naming practices. The experiment evaluated 120 arbitrarily selected open-source Java and C# classes from different programmers with respect to four naming conventions. The results indicate that Java and C# programs do not always follow naming conventions. However, Java developers are more attentive than C# developers in terms of following naming practices. A disturbing trend was found in variable and constant naming conventions, which were violated in most C# subjects. Moreover, there is a positive correlation between the number of violations found in a C# class and its size but a negative correlation in case of Java class. The findings are expected to contribute to the existing knowledge of the use of coding standards and source code quality. The paper also discusses the threats to the validity of the study and suggests open issues for future research.