English Abstract: | Android as a free open platform has become increasing popular and been widespread applied in mobile, tablet and other devices. An enormous amount of applications called Android apps have been developed so far. Based on the number of apps and developers, Android is a success. However, a great number of issues, such as apps in unequal quality, fragmentation phenomenon, have emerged along with its crazing development, enhancing the difficulty of developing and testing, as well as influencing the performance of Android apps and reducing the income of developers. Whether Android apps run fluency or not affects user experience directly. However, very few studies have analyzed the smooth performance and the frame rate related information. Besides, current tools which cannot obtain or record the frame rate across system and multiple applications cannot meet the demand of automated testing for Android apps in quantity. In this work, running fluency from the user and developer perspective is focused on, and various issues that Android apps does not run smoothly are introduced and analyzed, along with Android graphic system which draws UI. Then ARFluency – a tool to measure and test automatically across the system and multiple applications without modifying the source code of apps is presented, and an experiment to test apps from Google Play including 5 free top selling apps and AnTuTu Benchmark is conducted. After analyzing experimental data, I find that although these apps run fluency, they do all have some frames rendering problems. Also the frame rate may not describe the smooth performance accurately while it may qualitatively shows the GUI running fluency. |