Graphic User Interface Modelling and Testing Automation

A Graphical User Interface (GUI) is the most widely used method whereby information systems interact with users. According to ACM Computing Surveys, on average, more than 45% of software code in a software application is dedicated to the GUI. However, GUI testing is extremely expensive. In unit testing, 10,000 cases can often be automatically tested within a minute whereas, in GUI testing, 10,000 simple GUI test cases need more than 10 hours to complete. To facilitate GUI testing automation, the knowledge model representing the interaction between a user and a computer system is the core. The most advanced GUI testing model to date is the Event Flow Graph (EFG) model proposed by the team of Professor Atif M. Memon at the University of Maryland. The EFG model successfully enabled GUI testing automation for a range of applications. However, it has a number of flaws which prevent it from providing effective GUI testing. Firstly, the EFG model can only model knowledge for basic GUI test automation. Secondly, EFGs are not able to model events with variable follow-up event sets. Thirdly, test cases generation still involves tremendous manual work. This thesis effectively addresses the challenges of existing GUI testing methods and provides a unified solution to GUI testing automation. The three main contributions of this thesis are the proposal of the Graphic User Interface Testing Automation Model