Towards Autonomous Testing of GUI Applications

This whitepaper discusses the automation of test case design and implementation in the context of autonomous testing for graphical user interface (GUI) applications. It outlines the significance of automating test case creation to reduce costs associated with manual testing processes. The paper presents a proposed approach consisting of four phases: exploration, generalization, optimization, and validation. During exploration, the autonomous tester interacts with the application under test (AUT) to generate input data and log actions. The generalization phase involves estimating a state machine model based on collected traces, while optimization focuses on improving test cases. Validation assesses the correctness of the AUT, although it is limited to universal malfunctions. The paper also addresses challenges in achieving full autonomy in testing, such as unknown variables affecting the AUT's internal state and the complexity of certain actions. The findings were presented at the ESE Kongress 2023, contributing to the ongoing discourse on test automation methodologies.