| Knowledge brought to bear on design tasks is attributed to three sources: (a) problem solving strategies, (b) representation of physical objects and (c) inductive reasoning. In Part I an a priori model called Design Information Processing System (DIPS) is proposed. DIPS accounts for all three sources of knowledge and provides a general framework for the design activity. In Part II three empirical studies are done to calibrate the contents of each knowledge source. In the first study, design behaviors of architects are observed. In the second study, the structure of information retrieval from memory with architectural drawings is examined. In the third study, a framework for inductive reasoning in a drawing-interpretation task is explored. The results of these studies are represented through three data structures: (a) problem solving heuristics, (b) visual chunks and (c) if-then structures. A computer simulation model called Architectural Inference Marker (AIM) is designed and implemented in Part III, to test DIPS. AIM uses all three data structures to simulate subjects' drawing-interpretation behaviors. Implications of these findings for: (a) modeling design behavior, (b) identifying new research areas, (c) developing design aids and formalizing design education, are discussed. |
