基于随机性质的移动应用软件功能错误自动化检测技术

移动应用软件是一种基于图形界面操作的非传统软件，更新迭代快、用户基数大。功能错误是其最常见的一类软件错误，如特定功能突然失效、用户信息突然丢失、界面信息出错等。不少功能错误是在软件发布后才被发现，严重影响用户体验和忠实度，降低软件竞争力，重则对用户/软件公司造成不良后果。然而，目前仍然没有相关技术能自动化检测功能错误：(1)学界已有的测试技术局限于自动检测崩溃错误，无法察觉功能错误;(2)业界普遍以人工编写测试脚本验证功能正确。为此，本课题将通过如下三点研究内容，以Android软件为对象，提出一种基于随机性质的功能错误自动化检测理论、方法和技术：(1)研究功能错误的特征、表现形式和共性机理；(2)研究利用蜕变测试思想实现全自动功能错误检测；(3)设计一种轻量级功能测试规范语言，提高测试精准度和效率。本课题的研究对移动软件(以及其他图形化软件)的自动化功能错误检测具有重要的学术和应用价值。

Mobile apps are GUI-based event-driven software, which are fast-evolving and have huge number of users. Functional bugs are one of most common errors in mobile apps, e.g., unexpected functionality failures, lost user information, incorrect display information. Quite a number of functional bugs in mobile apps are only noticed post-deployment. They severely disrupt user experience and loyalty, negatively affect app success, and even cause severe loss for users and software companies. However, to the best of our knowledge, no existing automated testing techniques can fully-automatically detect functional bugs: (1) current automated testing techniques for mobile apps are limited to crash bugs; (2) manual testing is the most common practice in industry to validate functional correctness...Our research project aims to fill this gap by proposing an automated property-based functional testing technique for Android apps. In particular, we will investigate the following key research topics: (1) investigate the characteristics, manifestations and root causes of functional bugs in Android apps; (2) leverage the idea of metamorphic testing to realize an effective, automated functional testing technique for Android apps, and overcome the unique challenge due to lack of functional specifications; (3) design a lightweight domain-specific language for functional testing of Android apps, and thus improving testing precision and efficiency. The research results will benefit both academic and industrial communities, and shed light on automated functional testing for other GUI-based software.