A framework for Adaptive Capability Profiling

This thesis documents research providing improvements in the field of accessibility modelling, which will be of particular interest as computing becomes increasingly ubiquitous. It is argued that a new approach is required that takes into account the dynamic relationship between users, their technology (both hardware and software) and any additional Assistive Technologies (ATs) that may be required. In addition, the approach must find a balance between fidelity and transportability. A theoretical framework has been developed that is able to represent both users and technology in symmetrical (hierarchical) recursive profiles, using a vocabulary that moves from device-specific to device-agnostic capabilities. The research has resulted in the development of a single unified solution that is able to functionally assess the accessibility of interactions through the use of pattern matching between graph-based profiles. A self-efficacy study was also conducted, which identified the inability of older people to provide the data necessary to drive a system based on the framework. Subsequently, the ethical considerations surrounding the use of automated data collection agents were discussed and a mechanism for representing contextual information was also included. Finally, real user data was collected and processed using a practically implemented prototype to provide an evaluation of the approach. The thesis represents a contribution through its ability to both: (1) accommodate the collection of data from a wide variety of sources, and (2) support accessibility assessments at varying levels of abstraction in order to identify if/where assistance may be necessary. The resulting approach has contributed to a workpackage of the Sus-IT project, under the New Dynamics of Ageing (NDA) programme of research in the UK. It has also been presented to a W3C Research and Development Working Group symposium on User Modelling for Accessibility (UM4A). Finally, dissemination has been taken forward through its inclusion as an invited paper presented during a subsequent parallel session within the 8th International Conference on Universal Access in Human-Computer Interaction.