Building a tool for detecting code smells in Android application code

Abstract

In recent years, quality of the code behind the presentation layer has become increasingly important since a high number of applications are in maintenance mode. Maintaining complex applications is hard and modifying code that is difficult to understand may introduce new bugs. Poor design or implementation choices that contribute to technical debt are called code smells. Static analyzers are tools used to detect code smells and other vulnerabilities inside software applications. In this paper, we analyze static analysis tools and also develop an alternative, which tries to overcome the shortcomings of the previous tools by providing more code smell detection rules. As a result, we developed a plugin for SonarQube that analyzes applications written in Java programming language. Finally, we verified our tool by performing an empirical study on a selected corpus of applications.