Abstract

Empirical Modelling (EM) is an approach to the construction of interactive computer-based artefacts that embody construals of phenomena. The modelling activities underlying EM are associated with the identification of agents, dependencies and observables that reflect a commonsense way of understanding phenomena as experienced in everyday life. Model construction is guided by heuristic observation, interaction and experiment rather than by rational prescribed steps.

This thesis is a comprehensive study of the potential of EM as an approach to system development. EM rests on a philosophical foundation that is radically different from that underlying conventional approaches to system development. This is reflected in the reference to two perspectives on systems in the thesis title: before and beyond systems. The `before systems' perspective is concerned with the extensive preliminary activities that help to inform the system conception. These activities are supported by building models as prototypes to address the personal and interpersonal demands of system development. The `beyond systems' perspective is concerned with fulfilling the functions of a system without the circumscription characteristic of traditional systems. This has particular relevance for potential applications of EM to ubiquitous computing, where system conception is inseparably linked with system use in a situation.

This thesis seeks to consolidate and extend diverse concepts drawn from previous EM research within a unified framework: the Definitive Modelling Framework (DMF). The DMF supplies a suitable setting for both the cognitive and the collaborative aspects of system development in which the emphasis is on heuristic human problem solving and maintaining conceptual integrity in a system design.

Evaluations of software tools that support modelling within the DMF are conducted. The prospects for future tool development are extensively studied and illustrated by the construction of three new prototypical visual tools.

The research reported in this thesis provides a solid foundation for future research on applying EM to meet the challenges of system development in a modern computing context.