Multi-level Interaction in Parametric Design

Parametric design systems model a design as a constrained collection of schemata. Designers work in such systems at two levels: definition of schemata and constraints; and search within a schema collection for meaningful instances. Propagation-based systems yield efficient algorithms that are complete within their domain, require explicit specification of a directed acyclic constraint graph and allow relatively simple debugging strategies based on antecedents and consequents. The requirement to order constraints appears to be useful in expressing specific designer intentions and in disambiguating interaction. A key feature of such systems in practice appears to be a need for multiple views onto the constraint model and simultaneous interaction across views. We describe one multiple-view structure, its development and refinement through a large group of architecture practitioners and its realization in the system Generative Components. 1 Architectural practice and parametric design Conventional CAD systems focus design attention on the representation of the artifact being designed. Currently industry attention is on systems in which a designed artifact is represented parametrically, that is, the representation admits rapid change of design dimensions and structure. Parameterization increases complexity of both designer task and interface as designers must model not only the artifact being designed, but a conceptual structure that guides variation. Parameterization has both positive and negative task, outcome and perceptual consequences for designers. Positively, parameterization can enhance search for designs better adapted to context, can facilitate discovery of new forms and kinds of form-making, can reduce the time and effort required for change and reuse, and can yield better understandings of the conceptual structure of the artifact being designed. Negatively, parameterization may require additional effort, may increase complexity of local design decisions and increases the number of items to which attention must be paid in task completion. Parametric modeling has become the basis for most mechanical CAD systems and now is beginning to emerge as a tool for architectural design. While there is a general appreciation of the concepts and advantages of parametric modeling, application to projects at the scale and complexity of buildings raises important theoretical and practical issues. In a deep sense, parametric modelling is not new: building components have been adapted to context for centuries. What is new is the parallel development of fabrication technology that enables mass customization. Building components can be adapted to their context and parametric modelling can represent both context and adapted designs. In a design market partly driven by novelty, the resulting ability to envision and construct new architectural forms rewards firms having such expertise. There are relatively few such firms, most of which have had long experience and have built substantial reputations on distinctive form and construction. But many firms and students (future practitioners) are interested. The confluence of technology and interest appears as exploration in a new design space: architecture and its supporting technologies of parametric design and fabrication are experiencing both co-development and rapid change. Emblematic of this change is the SmartGeometry group. It comprises senior practitioners, a CAD system developer and academics. Its website leads with the declaration: Architecture is fundamentally about relationships. Many of those relationships are geometric in nature or find a geometric expression. The SmartGeometry group has been created in the belief that Computer Aided Design lends itself to capturing the geometric relationships that form the foundation of architecture....The group is dedicated to educating the construction professions in the new skills which will be required to use these new systems effectively....The group conducts a series of schools and seminars where this new technology is explored in the context of highly experienced professionals. (Minor verb tense changes made to the quotation.) (www.smartgeometry.org) To date, several such workshops have been conducted and workshop participants have continued to develop both designs and ideas for needed system support. The outcomes of the workshops and work that follows from them comprise designs, some intended for actual construction, some as explorations of design possibility. Figures 1 and 2 show work done by two workshop participants. SmartGeometry works, as it must, through using (and developing) systems that support design relationships. Much of its work parallels and is influencing the development over the last eight years of a specific parametric design system Generative Components (GC) by Bentley Systems, Inc. The primary designer and developer of GC is Robert Aish. At the time of writing, GC was not on the commercial market and its development path lay outside of the Bentley product process. Its design and development have been extensively affected by members of the SmartGeometry group, the professional members of which have acted largely in addition to their professional roles inside their respective firms. Academics and graduate students have been involved in design, review and trial workshops to an extent unusual for a corporate project. The existence of a motivated independent user community and a relatively open and relatively Fig. 1. Lars Hesselgren is a Senior Associate Partner and Director of R&D at KPF Architects. The images represent form explorations for a project currently (as of Spring 2005) under development at KPF. (by permission of the author)