B3 – Architectures of Algorithmic Order
From left to right: 1- Turing bomb at Bletchley Park (1945); 2- John McCarthy playing chess with a computer (1966); 3- ELIZA conversation program (1966); 4- URBAN 5 project (1973); 5&6- Seroussi pavilion proposal, Xefirotarch and biothing (2006); 7- Living Morphologies, supermanoeuvre (2009); 8- Cliff House, kokkugia (2012).

SECTION B – Order as Knowledge

B3 – Architectures of Algorithmic Order

Nadja Gaudillière , Oliver Tessmann

Department of Architecture at the Technical University of Darmstadt

The computational field in architecture has been developing since the 1960s, experimenting with digital tools and, more specifically, with programming languages and procedural processes in order to produce architectural designs. The tools, theories and practices associated with the field, initially developed in their majority in academic institutions, then gradually permeated the construction industry and architecture firms. This expansion of computational practices has gained momentum in the past years, and a wave of democratization has recently begun to occur, with an exponential number of practitioners resorting to such tools. Among those practitioners, a dual dynamic of appropriation and rationalization is at play. Appropriation is characterized in this context by the hijacking of digital tools originating in other fields of practices, and by the hijacking of computational rules of algorithmic typologies to which the practitioners resort. Architectural expertise is then implemented in custom algorithmic models, enabling a translation of the architect’s tacit knowledge into procedural structures and rules that are executable by a computer. Rationalization, on the other hand, favors material and industrial constraints, quantitative evaluations and computation models produced by scientific disciplines related to building design, known as being “objective”.

The field is characterized in particular by the capacity of its practitioners to develop their own design tools, from computer programs and plug-ins to robotic fabrication devices. In most of the recently developed tools, used not only by their developers but also by other users, a rationalist approach tends to be favored. This is apparent both in the socio-historical network of practitioners and in the tools’ structures and interfaces, to the point where the rationalist approach is embedded in black box tools. By being integrated into tools that privilege an easy, fast-tracked resorting to complex algorithmic models, rationalization becomes a technical and epistemological bias in the democratization process rather than a conscious choice by practitioners.

While a simplification of the use is necessary for the democratization of computational design tools, should it be to the detriment of the specificities of the architectural practice, as well as to the detriment of the existence of a multiplicity of subjective approaches to architectural design? The project aims to explore this question by developing tools and methods that allow for the democratization of computational design while preserving the core understanding of the tools and the models they are built on by users and reinforcing the capacity to implement architectural expertise in computational models. The research focuses on three key areas: questioning the nature of architectural expertise by studying tacit and explicit knowledge mobilization in algorithmic design; questioning the role of interfaces in our understanding of computational design tools; and questioning the possibility of modelling for unexplicit knowledge structures.