ACADIA 2012: Synthetic Digital Ecologies

Thérèse Tierney

The Association for Computer Aided Design in Architecture (ACADIA) convened late last fall at California College of the Arts (CCA) in San Francisco for its annual four-day conference, organized this year by CCA’s Jason Kelly Johnson.[1] It headlined an impressive list of international speakers, including Manuel DeLanda, Saul Griffith of otherlab, Greg Lynn, and Achim Menges.

Few would guess that ACADIA was founded in 1981, at a time when computational methodologies were isolated and esoteric practices in architecture. That has changed: the first day of the conference focused on practica, with software positioned at the forefront of the debate. Issues related to analytical feedback between data and design parameters were critical topics, represented by three principal methodologies: 1. Bespoke software designed by architects for specific applications; 2. Multiple proprietary software platforms for diverse applications; 3. One multifunctional inclusive proprietary software platform for diverse applications.

As a demonstration of the first position, UNStudio’s Gustav Fagerström discussed softBIM, an ad hoc open framework for design. softBIM associates data for design purposes with 3D models through custom made tools. Their experimental approach has two objectives: first, to gain increased control during the schematic stage of design, and second, to advance a specific open-source methodology. softBIM integrates commonly accessible platforms–in this case, Excel coupled with Rhino–yielding a flexible, lightweight system that is able to adjust to design ideas quickly while simultaneously allowing for continual feedback from the parametric data shaping the design.[2]

The second position, and probably the most widespread, was voiced by Reg Prentice from Gensler Architects during the Softwares Futures panel. The firm enlists diverse software platforms depending upon the stage of design development and designer preference. Such a heterogeneous approach is by all accounts the most pragmatic, and is already practiced by some firms as a matter of necessity. The concern, however, with this multi-faceted approach is twofold: software interchangeability, and that multiple proprietary platforms are commensurate with multiple annual licensing fees, potentially limiting its widespread applicability.  

The third position attempts to resolve these previous concerns. It was represented by Autodesk’s Robert Aish, who has been developing DesignScript, a multi-criteria design optimization program. The inclusive software package integrates a form-finding solver called SMART Form (developed in conjunction with Buro Happold), Ecotect Building Performance Application, and Robot Structural Analysis (RSA).[3] Aish has combined Generative Components transaction files (from Aish’s earlier work at Bentley Systems) with dataflow programming techniques.[4] Autodesk is betting on this approach to deliver one universal language to many different parametric interfaces. However, the real draw of DesignScript lies with two particular products: AutoCAD and Revit. Since AutoCAD and Revit both belong to Autodesk, compiling scripts into Revit families would eliminate the current cross-platform practice of baking Grasshopper or Digital Project models and importing them as static geometry to be manipulated by Revit.[5] Instead, a designer would be able to open the DesignScript model in Revit and associate it directly with the geometry in Revit, solving the software interchangeability problem discussed above. While the program is still under development, for those who like muscle software, DesignScript may make parametric form generation as common as BIM.

On Saturday and Sunday, the conference shifted its focus to academica with peer reviewed research papers. One of the most provocative was MIT’s Skylar Tibbits, who is on a mission to realize John Von Neumann and Stan Ulam’s concept of self-replicating automata, interpreted here as replication in natural systems. He is collaborating with a capable team that includes Arthur Olson of the Molecular Graphics Laboratory at the Scripps Research Institute and industrial designer Martin Seymour. Material computation has implications for replicating genetic sequences which determine self-folding proteins, essentially translating between genetic code and structure. Tibbits is applying similar abstractions to material systems through self-organizing and self-constructing components. While his research is still under development, it has significant potential for self-constructing assemblies. Embedding computational logic within physical material may offer new methodologies for future construction processes.

Also compelling was Plis/Replis, a sound project commissioned for Vranken Pommery Monopole, a champagne producer in Reims, France. The installation realized complex folding techniques through contemporary computer-aided design and manufacturing processes. Architect Hyoung-Gul Kook, along with artists Ali Momeni and Robin Meier designed, fabricated, assembled, and suspended a 345 cubic-meter structure built from 285 flat sheets of aluminum/polyethylene composite, robotically folded 2,535 times. The pavilion functions as a loud speaker, acoustically amplifying the sound of champagne bubbles placed beneath the folded shell, literally surrounding listeners in an effervescent environment. By engaging the physical senses, Pls/Replis expands architectural possibilities beyond experimental fabrication to shape space through audio effect.

In addition to presentations, the conference included a regional exhibition of material investigations by firms such as Future Cities Lab, Ogrydziak Prillinger Architects, IwamotoScott, and Rael San Fratello Architects, among others. There were experimental prototypes, also part of the exhibit, such as Andrew Kudless’ P_BALL, and Terri-Form, a collaborative project from the Architectural Association.[6] Special recognition went to Christine Yogiaman and Ken Tracy of Washington University, the award winners of Applied: Research through Fabrication, for their innovative cast concrete branching system.

For over a quarter of a century, ACADIA has been dedicated to advancing computational design methodologies in architecture. Today a computational approach necessitates a synthetic approach inclusive of diverse ideas and disciplines inspiring new ways of thinking and making. While questions will inevitably surface during any discussion of experimental methodologies, the research and projects presented at ACADIA emphasize that form-making is about understanding the intrinsic logic of form and structure. In that sense, what the philosopher Manuel DeLanda was alluding to in his keynote address was not a new type of essentialism, but rather, that material properties, expressed mathematically, underlay all apparent form–in nature as well as in designed artifices.[7]  For DeLanda (and Deleuze as well), design is not something imposed from without, but emerges from within matter as a topology.[8] Thus geometry is an underlying abstraction forging a connection between computation and formal architectural creativity. That dialectic of form, whether in architecture or in nature, involves understanding fundamental processes and the ability to transfer those processes into formal prototypes–by establishing a dialog between various contexts, parameters and constraints. DeLanda is implying that any investigation into form, whether nanostructures or skin scaffolding or any of the numerous ways and programs that computational geometry is employed today, are based on a generative logic of mathematical principles.  That is not to oversimplify, since there are other mitigating factors hidden within natural processes–but nature and computation are not so far apart as once thought.[9]