Supporting interactive system testing with interaction sequences

Abstract

Despite extensive research into the modelling and testing of interactive systems, existing strategies do not adequately cover all parts of an interactive system. These existing strategies model and test either the functional or interactive components of an interactive system separately, however, issues may arise where these components intersect. Therefore, further investigation into the modelling and testing of this intersection is required. Interaction sequences are a series of steps a user can take to complete a specific task or to arbitrarily explore an interactive system. In this research interaction sequences are used as an abstraction of the interactive system to inform a model-based testing approach using lightweight formal methods. Interaction sequences provide an abstract view of the point at which the functional and interactive components intersect, and as a result also provide a good starting point for investigation into the modelling and testing of this area. Interaction sequences are applicable to all types of interactive systems irrespective of the type of interaction, therefore modelling and testing approaches using this abstraction are also applicable to all types of interactive systems. In this thesis the findings of our investigation into modelling and testing using interaction sequences are presented. We describe formalisation of interaction sequences and modelling of these sequences using Finite State Automata (FSA). We introduce the self-containment property and show how this is used to control the size and state space of FSA. We demonstrate simulating interaction sequences and discuss how these models can be applied within both model checking and testing techniques. Lastly, we present a new approach for generating tests from interaction sequences and their associated models.