The software market is becoming more challenging as demands for faster time-to-market and higher software quality continue to grow. These challenges are embedded in all areas of Software Engineering, including Verification and Validation where they are proposed as solvable with automated testing. However, most automated testing techniques focus on low system level testing and are not suitable for high level tests, i.e. System and Acceptance tests, leaving industrial needs for test automation unfulfilled. In this paper we present a research plan to evaluate a novel automated testing technique, called visual GUI testing, based on image recognition algorithms and scripts that interact through the system GUI to automate complex scenario based tests. The technique has been evaluated at the company Saab AB where industrial, safety critical, scenario based, test cases were automated showing the industrial applicability of the technique. However, many factors are still unknown regarding the techniques industrial applicability, i.e. script maintenance costs, usability and learn ability, etc. Our research aims to uncover these unknown factors with the final research goal to show that visual GUI testing is a viable and cost-effective technique that will fill the gap in industry for a cost-effective, simple, robust, high-level test automation technique.
展开▼
