Design: Shape & Structure (decomposition / description)

Shortcomings in the cognate body of knowledge that underpins the discipline of engineering design result in slow incremental progress based on a series of ad hoc, idiosyncratic solutions with few if any high impact breakthroughs. This is problematic because engineering design is a critical early stage of the innovation processes that deliver new products to markets where societal challenges are addressed and wealth generated. This white paper reports results of a meeting that brought together a group of international research leaders. The working title for the meeting was "Design: Shape and Structure (decomposition/ description)" and the participating disciplines included architecture, computing, engineering and industrial design, manufacturing, and philosophy. The meeting took an existing research area, one characterised by experience-based frameworks and approaches, and explored the feasibility of establishing an integrated theory that draws together design with grammatical and mereological theories of shape and structure. Applying such theories to the synthesis of new designs could elevate the work of a designer above the detailed nuances of individual designers without diminishing their creativity and ingenuity. Benefits would include a well-founded theoretical and computational infrastructure on which the next generation of design systems could be built. The discussions focussed on aspects of engineering design that such an integrated theory might cover, how it might be used and who would benefit. With respect to coverage, three areas were identified: design process, functional aspects of designed artefacts and physical aspects of designed artefacts. The long-term impact of an integrated theory for engineering design would include coherent underlying principles for design definition and synthesis that could be used in design practice and education and as foundations for the development of future design systems. This, in turn, could lead to step-change improvements in the performance of the innovation processes that depend on effective and efficient engineering design processes, so resulting in the faster delivery of improved products to market using fewer (eg, financial and natural) resources. This white paper reports discussions and key conclusions from the meeting. The importance of distinguishing between design as a noun and designing as a verb was recognised. Given the broad range of areas and sectors such a theory would need to cover, a collection of related theories was regarded as more feasible and valuable. When considering design processes, a key requirement for any design theory lies in supporting zig-zagging processes between physical (for 3D products) and functional aspects of a design. There are well-established mechanisms for the definition of 3D products, developed through the development of 3D CAD systems and associated applications, but further work is needed to create a similar depth of understanding for functional aspects. ‘Capacity’ was put forward as a general concept that covers both function and behaviour. For designing as an activity, the critical unmet need in today’s computational design support systems lies in support for the ambiguity that designers need during creative episodes. Two areas for further development were identified: embedding and the need for computational tools that can operate on visual objects (so-called “visual machines”). In the shorter term, general purpose implementations of embedding will support design ambiguity by allowing visual and other emergent objects to be superimposed on existing design definitions. In the longer term, visual machines will be designed to support ambiguity by operating visually rather than symbolically. A roadmap for the development of these mechanisms is provided towards the end of this white paper.