(A) user experience modeling and evaluation framework for mobile usability

Abstract

The worldwide mobile software market has grown dramatically since feature phones became popular in the early 1990s. In practice, mobile usability - which can be defined for resource-constrained devices in two ways, i.e., through the User eXperience (UX) and through the User Interface (UI) - has been regarded as the key to gaining superiority in terms of both market share and customer loyalty. Unfortunately, de facto standards for software design and the development process, such as Unified Modeling Language (UML) and the Rational Unified Process (RUP), do not appear to promote mobile usability in a systematic manner in practice. This dissertation proposes a systematic and generalizable approach to modeling and evaluating mobile usability, herein treating it as a first-class software quality from the perspective of software engineering. We devise a UX evaluation framework for mobile usability, which we call the UX Evaluation Framework (UEF). The UX is specified by inter-scene interactions between users and the terminals of software products using Extended Menu Navigation Viewpoint (EMNV). A model checker, NuSMV, is then utilized to observe whether the EMNV diagram meets a set of given UX properties. Importantly, the Analysis & Design discipline of the RUP is extended to support the co-design of the UX and UI, so that the major roles, activities and artifacts in the UX and UI can be explicitly monitored and controlled by stakeholders. Through case studies, we demonstrate that UEF works properly to treat software products that prioritize mobile usability. Consequently, UEF plays a key role in filling the gap between the two research disciplines to address usability: Software Engineering and Human-Computer Interactions. UEF can help stakeholders conduct cost-effective and early UX evaluations in the design phase by suggesting the process of UX design, the EMNV diagram (a UX description model via inter-scene interactions), and the formal evaluation method involving the use of EMNV diagram.