Ideas
Computational Design seeks to understand design and to use computers to do so. Our vision is to create compelling computing aids that complement the capabilities of designers in general and flexible ways.
Achieving these aims requires a multi-disciplinary basis. In understanding design we proceeded from the work of others, for example Prof. H. Simon (late of Carnegie Mellon) and Prof. A. Newell (late of Carnegie Mellon), who set out models of human thought that describe higher level cognition in information processing terms. Such models are a ready structure onto which computational extensions can be grafted. In recent years, We use ethnographic methods to understand human thought and action at larger time scales and with greater ecological validity than typically achieved with information processing models. We have been influenced by those, for example, Prof. R. Coyne (Edinburgh), who claim that design is a phenomenon only partially apprehended by any single model. In understanding how computing applies to design, we proceed partly from formal models of design generation and representation developed by, for example Prof. U. Flemming (Carnegie Mellon), Prof. R. Krishnamurti (Carnegie Mellon), Prof. William Mitchell (MIT), and Prof. G.Stiny (MIT). These models posit a space of designs that can be traced out by operators that generate designs from other designs. To use such models requires the representation of designs, especially their geometry. Thus the large field of geometric and solids modelling are important sources for us.
We do both basic and applied research. There is much yet to be understood about design and computer models that might support it. However, such models fail a test of falsifiability if they remain unimplemented. In addition, questions of the appropriate human-computer interfaces cannot be addressed in the abstract—progress here relies on well-founded implementations. We see our research mission as building basic theory, and constructing, testing, using and disseminating representations, algorithms and implementations, both to develop and test theory and as tool prototypes in their own right.
What is especially unusual about our work is its implementation emphasis. It is clear to us that high-quality, tested, prototype implementations of design generation mechanisms are essential to future theoretical progress in the field and to its application outside of academia. Knowledge remains the product of our work, but implementations and their use are necessary prerequisites.