Design spaces: The explicit representation of spaces of alternatives

This Special Issue is in an exceptional format. Starting with an invited Keynote by Rob Woodbury and Andrew Burrow, the rest of the issue is taken up by invited papers that are responses to Woodbury and Burrow’s paper, written by selected authorities on design. A final short section is devoted to Woodbury and Burrow’s reactions to the responses. Rob Woodbury and Andrew Burrow set the stage in their paper “Whither design space?” The exploration of design spaces is a long-standing focus in computational design research. Design space exploration is the idea that computers can be used to help designers by representing many designs, organizing them in a network structure that forms the space, and by assisting designers to explore this space: that is, to make new designs and to move among previously discovered designs in the network. Three main areas of research into design space exploration can be distinguished. The first area of research concerns accounts of designer action and aims to reproduce and extend the behavior of designers. It is based on the premise that exploration is a good model for designer action. The second area of research aims to develop strategies and tools that amplify designer action in exploration. The third area of research concerns the discovery and development of computational structures to support exploration, including representations of the design space itself. Woodbury and Burrow specifically focus on computational access to the design space and the implications of having a design space representation in reference to the premise that exploration is a good model for designer action. Possible structures for a design space are conditioned by models of exploration behavior, by choices of strategies for amplifying designer action, and by the limits imposed by both computation itself and our knowledge of it. Formalisms for design space exploration must simultaneously accord with designer action, implement a useful amplification strategy, and be both formalizable and computationally tractable. Woodbury and Burrow raise, and answer, a number of questions, for example, what defines a good representation? Are design rules or, instead, design operators, the appropriate encoding mechanism for design moves in the design space? What is the role of the explicit design space, that is, the part of the design space the designer has previously visited, and what is the role of trajectories of design moves in design exploration? These questions, and others, form the basis for a discussion that can serve as a stepping stone for future research into design spaces. Ömer Akın distinguishes the paper by Woodbury and Burrow from previous efforts at exploring and studying the design space as taking an axe to an old growth trunk, compared to whittling away at the bark. Even then, the trunk remains massive and vast. In “The Whittled Design Space,” Akın examines Woodbury and Burrow’s paper according to four criteria: completeness, discrimination, alternative approaches, and combining exploration in different problem domains. With respect to the latter criterion, the author suggests the structure–function–behavior model as an analogy for the central parameters of the search space paradigm. He concludes with a number of questions both for this Special Issue and for future research. Ulrich Flemming indicates having no fundamental difficulties with Woodbury and Burrow’s paper, apart from some quibbles here and there. In “Yes, and by the Way . . . ,” Flemming picks up on the issue raised by Woodbury and Burrow of taking the specific domain representation too seriously, in particular, confusing the knowledge level with the underlying symbol level. He suggests that “mixedinitiative” design support systems need a symbol level consisting of a distinct task layer on top of the computational layer where the computational power of the system rests. The author also expresses unease with some details of WoodReprint requests to: Rudi Stouffs, Chair for Technical Design and Informatics, Faculty of Architecture, Delft University of Technology, Berlageweg 1, 2628 CR Delft, The Netherlands. E-mail: [email protected] Artificial Intelligence for Engineering Design, Analysis and Manufacturing ~2006!, 20, 61–62. Printed in the USA. Copyright © 2006 Cambridge University Press 0890-0604006 $16.00 DOI: 10.10170S0890060406060045