Complicating the ‘logic of design’

A ‘logic of design’ was outlined by March in 1976. Since then it has gained wide acceptance among designers and recently has been used in several knowledge-based systems of design. This paper makes a suggestion for extending the model by taking into consideration the logical nature of the laws in the argument form. This focus on the logical nature of laws brings out interesting and subtle differences in the inference process which are lost in the current analysis. It also results in some interesting complications. It means that it is no longer adequate to talk of the three phases of design — performance prediction, knowledge acquisition, and design generation — in the simple categories of deduction, induction and abduction, respectively. In particular, it means that performance prediction is not necessarily (or even usually) a deductive inference.