Detection of Unused Native Methods code smells in Multi-Language Systems

It is common practice in software development to use multiple programming languages within a single project, known as multilingual development. While it offers substantial advantages like legacy integration and performance optimization, integrating code from various languages can introduce code smells that adversely affect the overall readability, maintainability, and performance of these systems. These code smells are called multi-language code smells. Unused Native Method declaration and Unused Native Method Implementation are two of the multi-language code smells that are related to unused code. These code smells unnecessarily increase the volume of the codebase, which leads to difficulties in understanding and maintaining the code. Since these smells negatively impact the non-functional requirement of a software system, they must be detected and refactored. The detection approach proposed in the literature had many limitations, which is why the results presented in the literature about the prevalence of these two smells were incorrect. This paper addresses the identified limitations and assesses the existence of these two smells in open-source projects. Our results show that 9.34% of JNI files are affected by Unused Native Method Declaration code smell, and 46.18% of JNI files are affected by Unused Native Method Implementation code smell, which confirms the existence of these two code smells in open source multi-language systems.