A Design Framework for Pervasive Computing Systems

The question this thesis addresses is: How can we design pervasive systems? Designing a system goes much beyond giving building instructions. To design a system, we would like to be able to relate our sys- tem to other existing systems. We would like to be able to have grounds on which we can make design choices amongst various possibilities. We would like to be able to learn from existing systems, and thus improve future systems. Finally, we would like to be able to prescribe the required technology, thus push- ing the development of technology along the line of satisfying actual needs.

To answer our question, we build on the established HCI tenet that there are three dimensions to situa- tions where humans and computers, people and technology, come in contact: the user, the task and the domain. In this work, we explain why the notions of user, task, and domain are not sufficient to help design the grand vision of pervasive or ubiquitous computing. The concepts we propose instead are citi- zen, sphere, and space respectively. These three elements form our design framework, based on which we have developed a design tool and method. Our design tool can be used to model and represent pervasive systems, evaluate the potential of privacy breaching, indicate situations where physical interaction with the system is not possible, and inform the designer of situations where cognitive overload could happen. Cou- pled with our method for inspecting such problems, we show how design choices can be explored, design alternatives evaluated and compared.

We illustrate the applicability of our ideas on several levels. First we apply our ideas to the implementation of a gestural interaction technique. Then we draw on the results of an ethnographic study of the A&E department of a London hospital to propose design solutions. Finally we look at the city of Bath, where we apply our ideas and framework to generate design recommendations.