Department of Computer Science & Engineering
The Hong Kong University of Science and Technology
Lili is a PhD candidate in Computer Science and Engineering under the supervision of Professor Shing-Chi Cheung at the Hong Kong University of Science and Technology (HKUST). She received her BSc degree at Nanjing University in 2015. Her research interests include software analytics, testing, and mining code repositories. Lili's recent studies focus on common problems in Android apps. In her first piece of work, Lili has conducted an empirical study to understand fragmentation-induced compatibility issues. She further proposed a static analysis technique, FicFinder, to detect potential compatibility issues in Android apps. This study was awarded an ACM SIGSOFT Distinguished Paper award and has received great interests from both research and industrial communities. Currently, Lili is working on extensions of this study to gain a more comprehensive understanding of fragmentation-induced compatibility issues in Android apps. This will also be the focus of her PhD thesis.Characterizing and Detecting Fragmentation-Induced Compatibility Issues in Android Apps
Android ecosystem is heavily fragmented. The numerous combinations of different device models and OS versions make it impossible for Android app developers to exhaustively test their apps. As a result, various compatibility issues arise. Such fragmentation-induced compatibility issues (FIC issues) have been well-recognized as a prominent challenge in Android app development. However, little is known on the characteristics of these FIC issues. To bridge the gap, we conducted an empirical study on 191 real-world compatibility issues collected from popular open-source Android apps. Our study characterized the symptoms and root causes of compatibility issues, and disclosed that FIC issues exhibit common patterns. With these findings, we proposed a technique, FicFinder, to automatically detect compatibility issues in Android apps based on issues patterns manually extracted from our empirical study. Since FIC issues are evolving as new Android versions and devices are released, manually extracted issue patterns can eventually get outdated. To address this problem, we aim to develop a novel framework that combines program analysis and data mining techniques to automatically learn FIC issue patterns from large corpora of Android apps. The learned patterns can be fed into FicFinder to automatically detect potential FIC issues. This can significantly reduce the search space for FIC issues and benefit the Android development community. To evaluate our proposed framework, we leveraged the mined patterns to detect unknown compatibility issues in open-source Android apps. FicFinder has revealed 63 previously-unknown FIC issues and 22 of them have already been fixed by the app developers.