Principles of design for a multiple viewpoint problem formulation support system

The world is filled with multiple views. There are multiple political philosophies, multiple economic models, multiple scientific theories and multiple management science models. This research explores the architecture of a computer system that creates, displays and manages multiple views. Three basic questions are answered: "What are the characteristics of a multiple viewpoint system?" "Can a multiple viewpoint computer system be constructed?" and "What design principles guide the development of a multiple viewpoint system?"

The answers to these questions are discovered using two approaches. The analytic approach uses aspects of set theory and logic to specify system characteristics. This approach also leads to some general design principles, including the principle of viewpoint normalization. The second approach is system prototyping. In particular, a multiple viewpoint system, DOVE, is created which supports management problem formulation. This system is capable of creating and managing multidimensional financial models, causal models, Forrester flow models, goal hierarchy models, linear programming models and many other types of views. This second approach leads to more detailed design principles concerning the architecture of frames and rules.

This dissertation develops the basic characteristics of a multiple viewpoint system, demonstrates that such a system can be constructed and illustrates principles which can guide the future development of a multiple viewpoint system. It contributes to the information systems field since past research has focused almost exclusively on the design of systems which promote a single point of view.