Towards the systematic assessment and design of adaptive user interfaces.

摘要

Adaptive user interfaces have been suggested as an alternative to help users deal with information overload and the complexity of software. However, experimental studies have questioned their effectiveness, while the introduction of adaptation techniques in commercial applications has had limited success. The negative past experience has caused skepticism among researchers in Human-Computer Interaction, as adaptive behaviour has been linked to poor designs, that have violated well-established usability principles. Several researchers have suggested that adaptable user interfaces, which are interfaces adapted by the user rather than the system itself, should be considered as a better alternative.;This dissertation explores the design space of adaptation techniques in a range of user interfaces, identifies tradeoffs between adaptive and adaptable user interfaces and makes recommendations about their design and evaluation. In particular, it examines three different types of interfaces: hypermedia systems, lists, and hierarchical drop-down menus. New adaptation techniques are proposed for each interface. We explore both techniques that assist navigation and visual search, and techniques that facilitate movement and selection. We address limitations of existing techniques by exploring designs that merge automation with user control. The proposed designs are extensively evaluated. Specifically, the dissertation documents a total of seven user studies. Through our evaluation work, we explore the role of adaptation accuracy as a measure of the quality of automatic inference and decision-making, and evaluate its effect on the success of adaptation techniques. This allows us to investigate strengths and limitations of the techniques that we propose and come up with recommendations about the design and evaluation of adaptive user interfaces.;Overall, the dissertation has three major contributions: (1) the introduction of designs of hypermedia systems that combine automation and direct user manipulation; (2) the systematic treatment of accuracy as a way to assess the effectiveness of adaptation techniques; and (3) the systematic design and evaluation of bubbling menus, a new design of adaptive and adaptable menus.