Architecture stands at an inflection point. The confluence of advances in both computation and fabrication technologies offers architects the possibility of designing and constructing hitherto unimaginable forms.
With increases in processing power, the roughly triangulated geometries and simple blobs of the early 2000's have given way to the possibility of complex geometries at multiple scales with details approaching the threshold of human visibility. In parallel, advances in additive manufacturing technologies have put us at the verge of printing any form. Recent machines with print spaces of many cubic meters make it possible to print not only small architectural models, but full-scale architectural components. As a result, a form with a few million surfaces is as easy to print as a form with a few dozen.
For the first time, complexity is not an impediment to design and fabrication. Rather, it is an opportunity that is waiting to be explored. For years, it was information technology that constrained architects. Arguably, this relationship has reversed: it is now architects who are constraining the possibilities of information technology. This development raises the questions: How can we best explore the opportunities that information technology offers us? How can we understand the possibilities?
To truly exploit the possibilities, we can no longer draw by mouse in CAD programs. A single object with millions of unique facets would take years to draw. Neither can the new opportunities be fully exploited using parametric approaches, as these usually involve morphing existing geometries using control parameters, rather than creating geometries that are genuinely new.
What is needed is a more abstract and open-ended method: a computational approach. In computational design, parameters do not control the geometry directly. Rather, they control the operations of a time-based, predefined process that is itself transforming or generating geometry. These processes strike a delicate balance between the expected and the unexpected, between control and relinquishment. These design processes are deterministic – so as not to rely on randomness, but not necessarily entirely predictable. Instead, they have the power to surprise.
Once formulated, such a computational approach can be applied again and again. One no longer designs an object, but a process to generate objects. It is no longer necessary to successively refine a singular design, as one can work with many variants in parallel. These variants can be bred and cultivated into entire families of objects by combining and mutating their constituent process parameters.
A computational approach enables architecture to be embedded with an extraordinary degree of information. Structure and surface can exhibit hyper-resolution, with seemingly endless distinct formations. The processes can generate highly specific local conditions, while ensuring an overall coherency and continuity. As such, the resulting architecture does not lend itself to a visual reductionism. Rather, the procedures can devise truly surprising topographies and topologies that go far beyond what one could have traditionally conceived.