Android application development has attracted attention of many software developers as a mainstream software platform. Despite its interesting characteristics, there are a number of issues that result in developing applications with poor performance. Hence, it is very crucial to evaluate quality of Android applications, specially their robustness and responsiveness. Monkey testing is a traditional technique used for testing these aspects, and Google’s Android Studio, as the official Android IDE, is equipped with monkey testing facilities. In this paper, two experimental studies are performed to investigate the necessity of monkey testing on Android applications and also to evaluate the effectiveness of the default Android Studio monkey for revealing application failures. The results demonstrate that there is a strong negative correlation between the number of crashes detected by the monkey and the rating that users have assigned to the application. So, monkey testing for Android applications is indeed a requirement. Further, the results show that the default monkey is weak in terms of focusing on the target application, and hence, there is much room for improving monkey testing techniques for Android applications.

Keywords