The status of design theory research in the United States David G Ullman Department of Mechanical Engineering, Oregon State University, Corvallis, Oregon, 97331, USA
This paper contains a discussion of United States design theory research. It begins with a definition of a design theory and a brief history of design theory research. This is developed into the factors that affect the research direction and results. The paper concludes with a listing of the types of research currently being persued. Keywords: design theory research, USA
INTRODUCTION This paper gives a brief picture of the status of design theory research in the United States. As a basis for this discussion, an effort will be made in the next section to define what is meant by the term 'design theory'. This will be followed by a brief history of design theory research in the United States with emphasis on the factors that affect its direction. In the fourth section current research areas will be identified. Finally, in the fifth section expected future directions for research will be identified. When reading this paper it should be remembered that it represents one researcher's views and that this researcher, although interested in all areas of design, is primarily a mechanical engineer. Additionally, this paper will focus on the influences and direction of design theory research rather than try to relate a specific theory. This is because there is not yet (in the USA) a specific theory to describe. It is still evolving.
W H A T IS A ' D E S I G N THEORY'? Theories are developed to explain natural phenomena.
204
The act of designing is certainly a natural phenomenon so it seems reasonable that developing theories about the design process is a worthy goal. However, as will be shown, few US researchers are driven primarily by this goal. Before supporting this statement, the definition of a theory will be refined. A theory has three basic components; a set of variables that measure the phenomena, a cluster a models relating the variables to explain the phenomena and hypotheses for the application of the models to the real world. 1 A standard mechanics textbook is built on variables (e.g. F, m, a), contains models based on Newton's laws of motion and has guidance (hypotheses) on how to apply these to real problems. We rely on this material to explain basic mechanics to our students. Theories about design can have one of three different foci. First, they can be formulated to explain the designed object itself, i.e. how the object comes into being and evolves into a final, manufactured product. Second, theories can be formulated to explain how the designer(s) transforms an ill-defined problem into a fully described product. Finally, theories can be developed that attempt to explain the process of design, the interaction of the artifact and the people and/or computers in the design
0142-694X/91/04204-05 © 1991 Butterworth-Heinemann Ltd
DESIGN STUDIES
environment. Theories developed about these three foci are not independent. Dixon 2 has argued that design theories can be prescriptive (what should be done), cognitive (what is done) or computational (how to do it). This breakdown is also useful. Another aspect of a theory that will help in understanding the state of research in the United States is along what might be called the pragmatic dimension. At one end of this dimension is the philosophical study of design that attempts to give an idealized rationality to it. At the other is the sociological view that seeks to explain the interaction of social and technological issues, i.e. what actually occurs. Any effort to explain the design itself, the designer or the design process can rely on an ideal model, a social model or anywhere in between. A final consideration is a theory's breadth of application. This is really part of the hypotheses for how to apply the model, but it is treated as a separate item as it is of importance in understanding design theory. In developing theory about design, research can range from a single class of artifacts in a limited domain (e.g. a theory of gear design) or phase of the design process (e.g. a theory of embodiment), to a general theory that applies to the design of anything regardless of domain or phase. Based on the three descriptors of a design theory, its foci, pragmatism and breadth, the following section gives some history of design theory research in the United States.
A BRIEF HISTORY OF DESIGN THEORY IN THE U N I T E D STATES This section will attempt to piece together a history of the interest in and influences on design theory studies in the United States. The discussion will begin generally and then focus on the leadership of the National Science Foundation, consumer product industries and the Department of Defense on the directions of design theory research. It must be borne in mind that the research areas pursued are greatly influenced by these three funding sources. The section ends with another indicator of the interest in the study of design theory - professional society activity. In the USA, the term 'design theory' was virtually unheard of prior to the early 1980s. Around 1980 texts covering the structured design of computer codes 3 and the systematic design of VLSI systems4 were published. These texts focused on specific domains and gave an idealized process for the development of the products within the domains. They and other texts in the same disciplines have had considerable influence. The structured techniques have spawned industries that manufacture computer tools to support these models. During the same period there were a number of books published that presented theories for mechanical design. In the United States, Planning and creating successful engineering designs: managing the design process by Love 5 gave a view of the mechanical design process that formed
Vol 12 No 4 October 1991
a view between practitioner and manager. This book however, was never very widely disseminated. Books by Hubka 6 and by Pahl and Beitz7 were also published at this period. Hubka's book is oriented toward a philosophical theory of the design process and has not been widely studied in the US. The text by Pahl and Beitz presents a model of the design process that is a compendium of techniques and design axioms. This text was not widely read in the US until late in the decade when it became a reference for design methodologists. It has also found some use as a text in graduate courses. In September 1985 the National Science Foundation (NSF) sponsored a workshop for representatives from the disciplines of chemical, civil, electrical, industrial and mechanical engineering, and computer science. The goal of this workshop was to establish goals and priorities for NSF funding of research in design theory and methodology (Rabins et al.S). This workshop identified five research areas as making up the discipline of design theory and methodology • • • • •
conceptual design and innovation quantitative and systematic methods intelligent and knowledge-based systems information integration and management human interface issues in design
They further identified a series of criteria on which to base the evaluation of proposed research. First on this list was
Contribution to theory: Will research in this area add to the fundamental body of knowledge and provided new theoretical and generic insight into the design process? The second and third criteria were oriented toward relevance to practice and entrepreneurial potential. Thus the participants in the workshop saw the development of underlying, domain independent, theories as a primary research direction for the NSF. Based on the results of this workshop the NSF Design Theory and Methodology program office began operation in 1985. Its goals were to fund research in (1) scientific theories of design, (2) foundations for design environments, and (3) models of the design process. In February 1987 the program had its first meeting (Waldron9). The attendees ranged from design philosophers (who were truly looking for an ideal model of design), to the pragmatists looking for problems on which to apply their computer algorithms. The participants also ranged across all disciplines from the fine artists, architects and philosophers, to computer scientists and mechanical engineers. The meeting was not very successful as there was no agreed definition of 'design' and little unity on the approach to studying it. In many ways it was like a tower of Babel with few speaking the same language. Since that first meeting the NSF program has steadily drifted from its original goals to the point where its emphasis is now primarily focused on the design of mechanical objects with a strong emphasis on design for manufacturing. Thus, the breadth of research funded has
205
narrowed to focus on mechanical objects and it has become very utilitarian and unphilosophical. Although NSF is a major guiding influence on research in the United States, other equally strong influences are the consumer product industry and the defense industry. Leadership in the consumer products industry has come from companies in highly competitive fields such as office equipment, automobiles and farm equipment. This leadership has taken three paths - the adoption of techniques from Japan and Europe, the improvement of information management and the funding of research on design issues. Each of these paths will be discussed. In the late 1970s the United States began losing consumer product and automobile markets to the Japanese and Europeans. For example, in 1980 the Japanese could sell a photocopier for the same price as Xerox could manufacture one. One Xerox study to find out why they were no longer competitive focused on the cost of individual parts. By comparing plastic parts from their machines to parts that performed similar functions in Japanese and European machines, they found that the Japanese could produce a part for 50% less than American or European firms. They attributed the cost difference to three factors: materials cost 10% less; wages, tooling and processing costs were 15% less and; the remaining 25% (50% of the difference) was attributable to how the parts were designed. 1° Based on these results companies like Ford and Xerox adopted methods from the Japanese such as quality function deployment and Taguchi's robust design techniques. The second emphasis in the consumer product industry is a strong effort on product information management. These efforts have focused mainly on computerized data bases. Although not directed at the development of a design theory, they aid in the evolution of data representations which embody the variables on which theories are based. The third emphasis has been in funding research. The proportion of research funded by consumer product corporations is unknown, but is thought to be comparable to that funded by the NSF. Research funded by these industries is primarily focused on understanding the design of objects with little effort on the process or the designer. Additionally, a majority of this research is on a specific product or narrow class of products. The US Department of Defense (DOD) has also had a strong influence on the direction of design theory research. Their efforts are built around the concept of concurrent design. In a report published by the Institute for Defense Analysis (IDA), an advisory group to the DOD, concurrent design is defined as Concurrent engineering is a systematic approach to the integrated, concurrent design of products and their related processes, including manufacture and support. This approach is intended to cause the developers, from the outset, to consider all elements of the product life cycle from conception through disposal, including quality, cost, schedule and user requirements, n
206
As can be appreciated from this definition the DOD effort is very broad and very pragmatic. Beyond the influence of the NSF, the consumer products industry and the DOD, the professional societies reflect the interest in design theory development. In 1987 the American Society of Mechanical Engineering (ASME) Design Theory and Methodology Committee was formed in the Society's Design Division. Although other professional societies have groups interested in design theory, only this committee has actually sponsored conferences on the subject. 12"~3 The first conference contained 13 papers and the second 45. Both conferences were held in parallel with other Design Division conferences. The design theory and methodology sessions drew 50 to 100 participants to each session, a very strong showing. A third meeting is scheduled for September 1991 in Miami, and will be restricted to 50 papers in an effort to keep the quality high.
TECHNICAL AREAS OF DESIGN THEORY STUDIES There are very few researchers in the US who are trying to find a basic, philosophical, underlying science of design. This search for an idealized model has been overshadowed by the development of domain and phase dependent models as a result of the interests of the financial sponsors. One other factor that has greatly influenced the development of design theories in the United States is the strength of the computer industry. Virtually every engineer in the United States has access to a computer and most have at least a PC on their desk. Thus there is a great emphasis on the use of computers in the design process.When mechanical Computer Aided Design (CAD) systems were becoming popular in the late 1970s, they were marketed as actual aids to the design process. By the mid 1980s it was realized that they were really best suited for drafting and were of little help for design. Both the CAD industry and the user community realized that the primary users of the systems were draftspersons and that very few of them were used by engineers. This realization encouraged research on improving user interfaces, developing tools to support earlier stages of the design process and achieving more sophisticated data representations. Work on data representation has had a strong influence on the development of a basic vocabulary describing the design objects and the design process. Through developing ways of representing design information in computers,-researchers have been forced to develop underlying models of the design process. Developing these models and their subsequent testing, has lead researchers to refine the models and push toward an understanding of the basic building blocks of a design theory. The development of computer science techniques has also influenced the direction of design theory research. The computer science community developed techniques
DESIGN STUDIES
such as expert systems, object oriented programming, hyper-medias, and cheap computing power. There is a desire on the part of both the computer science and engineering communities to put these techniques to work to help support design activity. The use of these tools forces the development of models and thus a theory of design. Actual research on design theory can be categorized in a number of different ways. In a two-part paper titled 'A review of research in mechanical engineering design 14'15 the discussion was organized into six research areas listed below. The number of US research projects mentioned for each area is shown in brackets • • • •
descriptive models of the design process (10) prescriptive models for design (3) computer-based models of design processes (58) languages, representations, and environments for design (31) • analysis to support design decisions (22) • design for manufacturing and other life cycle issues such as reliability, service ability, etc. (35)
Clearly the bulk of the research is focused on computational modelling and the underlying vocabulary needed to support these models. The influence of consumer product and defence industries is also shown.
FUTURE DIRECTIONS
In a report 16 released in February 1991 (NRC 90), the Committee on Engineering Design Theory and Methodology of the National Research Council (NRC) recommended that research in the United States follow a ten point, national research agenda, outlined below (A) Developing scientific foundations for design models and methods • computer representations for in-progress designs • generating, organizing and generalizing design knowledge • synthesis: parametric, configuration, and conceptual design • tolerance synthesis (B) Creating and improving design support tools • designer-oriented computational prototyping, analysis and simulation tools • rapid physical prototyping • design for 'X' (C) Relating design to the business enterprise • quality/cost models • organization and communication models • innovation Since reports from the NRC are very influential, it is expected that future research will be focused in these areas. The development of a design theory is not explicitly listed as one of the areas. Rather, models will be developed in each of these areas that contribute to an evolving theory.
Vol 12 No 4 October 1991
CONCLUSIONS This paper has focused on two topics; the factors that are influencing the development of a design theory in the United States, and the directions that design theory research is taking. From this discussion the following conclusions can be drawn • design theory research is in the pre-theory stage. There is still a search for the basic vocabulary and building blocks of a theory • the theories developed in Europe are not well known nor broadly studied in the USA • there is more a focus on computer based tool development than on theory development. There is a general feeling that through understanding and developing models of the parts, the whole may be understood • there is no clear common design research community. Rather it is still discipline and viewpoint fragmented • there is weak leadership on design theory research from the National Science Foundation. There is some from the DOD built around the concept of concurrent design, however concurrent design does not have theory development as its goal • there is virtually no research from the philosophical view point. This is not to say that researchers do not step back from their work periodically and try to philosophize about how it fits into an overall theory of design These conclusions do not imply that a design theory is not evolving in the United States. It is on an ad-hoc basis with no clear focus or goal. As models are developed to solve the pieces of the design puzzle, a theory will evolve. This is still at least seven years away. If only we could use a sound design methodology to approach the problem of designing a t h e o r y . . .
REFERENCES
1 Giere, R N Explaining science: a cognitive approach, The University of Chicago Press, Chicago and London, (1988) Dixon, J R, 'On research methodologytowards a scientific theory of engineering design' Artificial Intelligence for Engineering Design, Analysis and Manufacturing (A I E DAM ) , Vol 1, No 3, 1988 Yordin, E and Constantine L Structured design fundamentals of a discipline of computer program and systems design, Prentice Hall, (1979)
4 Mead, C and Conway, L Introduction to VLSI systems, Addison Wesley, (1980) Love, S Planning and creating successful engineered designs: managing the design process. Advanced Professional Development Inc, Los Angeles, (1980) 6 Hubka, V Principles of engineering design, Butterworth Scientific, Borough Green, UK, (1982) 7 Pahl, G and Beitz, W Engineering design, The Design Council (1984) (original German text 1977 Springer Verlag)
207
8 Rabins, M, Ardayfio, D, Balzar, R et al., 'Design theory and methodology - a new discipline', Mechanical Engineering, Vol 108, No 8, 1986, pp 23-27 9 Waldron, M, 'Results from the NSF workshop on the design process' Proceeding from the meeting at the Claremont Hotel, Oakland, CA, February 8-10 1987 10 Ullman, D G The mechanical design process, McGraw Hill, (in publication, due 1992)
13 Rinderle, J R, Deszgn theo~ and methodology DTM 90. ASME DE-Vol 27, (1990) 14 Finger, S and Dixon, J R, A review of research in mechanical engineering design, part I: descriptive, prescriptive, and computer-based models of the design process', Research in Engineering Design, Vol 1, No 1, (1989) 51-67
11 Winner, R I et al, 'The role of concurrent engineering in weapons system acquisition', Institute for Defense Analysis, IDA Report R-338, Fairfax, VA, (December 1988)
15 Finger, S and Dixon, J R, A review of research in mechanical engineering design, part II: representations, analysis and design for the life cycle', Research in EngineeringDesign, Vol 1, No 2, (1989) 121-137
12 Elmaraghy, W H, Seering, W P and Ullman, D G, Design theory and methodology - DTM 89, ASME DE-Vol 17, (1989)
16 National Research Council, Improving engineering design: designing for competitive advantage, National Academy Press, Washington D.C., (1990)
208
DESIGN STUDIES