Supporting multimedia user interface design using mental models and representational expressiveness

This thesis addresses the problem of output media allocation in the design of multimedia user interfaces. The literature survey identifies a formal definition of the representational capabilities of different media.as important in this task. Equally important, though less prominent in the literature, is that the correct mental model of a domain is paramount for the successful completion of tasks. The thesis proposes an original linguistic and cognitive based descriptive framework, in two parts. The first part defines expressiveness, the amount of representational abstraction a medium provides over any domain. The second part describes how this expressiveness is linked to the mental models that media induce, and how this in turn affects task performance. It is postulated that the mental models induced by different media, will reflect the abstractive representation those media offer over the task domain. This must then be matched to the abstraction required by tasks to allow them to be effectively accomplished. A 34 subject experiment compares five media, of two levels of expressiveness, over a range of tasks, in a complex and dynamic domain. The results indicate that expressiveness may allow media to be matched more closely to tasks, if the mental models they are known to induce are considered. Finally, the thesis proposes a tentative framework for media allocation, and two example interfaces are designed using this framework. This framework is based on the matching of expressiveness to the abstraction of a domain required by tasks. The need for the methodology to take account of the user's cognitive capabilities is stressed, and the experimental results are seen as the beginning of this procedure.