Designing the Computational Image, Imagining Computational Design – An Excavation of Digital Form

“Designing the Computational Image, Imagining Computational Design” is an exhibition that excavates the foundation of computer-aided design and manufacturing (CAD/CAM) and weaves together several ‘origin stories’ for contemporary consideration. The show recently closed after a seven-week run at the Miller Gallery at Carnegie Mellon University in Pittsburgh, and CAN was fortunate enough to get a guided tour with curator Daniel Cardoso Llach as it was winding down. The exhibition offers an encyclopedic pre-history of CAD/CAM and its applications across architecture and design, and also underscores links between this early research and present-day computational design.

“The exhibition deals with the developments of the first annotations and inscriptions in software to drive numerically controlled machines. It documents the development of these first computer numerically controlled machines – on punched tape or magnetic tape – and how these projects aligned industrial and academic interests,” says Llach as our tour gets underway. These annotations and inscriptions did not emerge from benign curiosity – they were rooted in “more Air Force per dollar” postwar military doctrine; unsurprisingly, the earliest forays into computer manufacturing were fueled by Cold War anxiety. Extending out of this nationalistic starting point, “Designing the Computational Image, Imagining Computational Design” deploys photographs, models, films, and software reproductions to delineate the early years of computational design and connect them to the present.

↑ Software takes command: this vitrine collects ephemera from some of the first experiments in computer-aided manufacturing (above), Charles Eastman’s 1980 experiments in applying graph theory to organize space (lower left), A.P. Armit & A.G. Flutter’s CAM-produced metal shoe (lower middle)

Stepping into the exhibition time warps the viewer back to the 1950s. “A convergence between materials, machines, and geometry in the space of software is what this section tries to describe through the MIT Servomechanisms Laboratory and the CAD Group at Cambridge University, which were the two research centres that were pushing these technologies.” Llach explains how the MIT’s ‘Servo Lab’ shifted focus from missile guidance in WW2 to aeronautics manufacturing when engineer/entrepreneur John T. Parsons brought them on as partners as part of a massive air force contract his company received in 1948. Some highlights of this ‘software comes to matter’ section of the show: Donald F. Clements 1953 diagram that illustrates the subdivision of a 3D elliptical volume into discrete parts; 3D graphics by Douglas T. Ross visualising paths for a milling machine (calculated by a Whirlwind I computer); and triptych of images documenting the production of a shoe machined from metal alongside the computational model that yielded it.

“In order to communicate with machines they had to think of geometry as something that was carefully described, parameter-driven, and turn it into a symbolic language,” Llach says of the archival wireframe graphics. Next, we proceed into the ‘structured images’ section of the show, which introduce the computer’s analytical capabilities to representations. Things open up here and more clinical material like Lionel March and Philip Steadman’s volumetric analysis of the Seagram building (1971) and a montage of curvature studies across disciplines (1967) are arranged alongside more explicitly aesthetic experiments including Paul Pagnaro’s animation of aperiodic patterns (1975) and videos produced by Lillian Schwartz and Ken Knowlton at Bell Labs (1968). “I found material that I felt could be put in conversation with thes more applied engineering works,” Llach says of the latter, playful works. “The structural image of these computational representations is not put in the service of engineering efficiency but more plastic visual explorations.”

Llach has been both meticulous and diligent in sourcing material, and the show draws from several archives including multiple MIT collections, the Computer History Museum, Cambridge CAD Group, as well as several personal collections. Given the depth of this research, we were not surprised to learn the show was informed by Llach’s dissertation – it yielded a book entitled Builders of the Vision, published by Routledge – as the show paints a vivid picture of the outputs and ambitions of proto-computational designers. And while this thoroughness is truly commendable, where “Designing the Computational Image, Imagining Computational Design” really shines is in two interactive reconstructions of important software created as part of its ‘interaction and intelligence’ selection of works: Steven A. Coons’ Coons Patch and Ivan Sutherland’s Sketchpad.

“The Coons Patch is a technique for parametric interpolation. It allows you to take four bounding curves – defined mathematically – and then calculate a surface along those curves.” Llach explains how Coons, a self-taught mathematician and “strange character” developed this technique to create hard-to-describe geometries that tangibly impacted aerospace manufacturing. The reconstruction allows you to manipulate control points on four lines that bound a surface with complex curvature, and (quite intuitively) move vertexes to generate new iterations of the surface; anyone who has used 3D modelling software like Rhino will be intimately familiar with these mechanics – and here they are in their ur-form.

Engaging the reconstruction of Sketchpad was similarly illuminating. One of the pillars of computer graphics and human-computer interaction, Sutherland’s 1962 presciently offered a GUI/light pen interface that allowed a user to draw lines and construct shapes and manipulate them with ‘drag and drop’ functionality as well as copy, paste, snap etc. – interaction paradigms that are all commonplace now. Llach on its enduring influence: “the drawing is no longer a pictorial representation of a shape but a structure – something that has its own mechanics … amazingly it encodes the basics of parametric design: associative models with constraints.” Llach worked with CMU grad student Scott Donaldson to rebuild the software and notes that in this case their ambition extended beyond what happened onscreen. “The reconstruction is not just of the graphics – we tried to reconstruct some of the ergonomics. It’s not a precise replica but it tries to evoke what it meant to sit down at Sutherland’s workstation.”

↑ Get contemporary: the exhibition presents a range of recent works to complement the historical survey, this includes generative (architectural-ish) drawings by Andrew Heumann and Dana Cupkova (lower left) and en elegant drawing transcription series by Carl Lostritto (lower centre left)

“Designing the Computational Image, Imagining Computational Design” could easily have been presented with all the enthusiasm of a dust-caked history lesson or (worse yet) relentlessly capitalize on nostalgia. Not only does it avoid these pitfalls, it bolsters its resonance with present-day computational design by presenting works by a selection of contemporary artists that are interrogating digital form. Beyond building bridges that span several generations of practice, something quite magical happens when you have a modular painting by George Stiny and experimental video by Lllian Schwartz in the same space as Zach Lieberman’s post-Instagram geometric studies, Joseph Choma’s robotically carved limestone sculpture, and a prescient interactive sketch by Golan Levin (originally built in Java, reconstituted here in iOS).

Before CAD (and Processing and contemporary creative coding frameworks) there was researchers like Sutherland and Coons; before Frank Gehry there was the aerospace industry; before the MIT Media Lab, there was the Servo Lab and the Architecture Machine Group; before sub-$500 desktop fabrication there were decades of clunky CAM experiments; “Designing the Computational Image, Imagining Computational Design” expertly uncovers the superstructures that underpin contemporary digital design – from architecture to software art – and does so in a way that is both accessible and engaging. At one point Llach makes an insightful observation about the ‘structured images’ portion of the show that encapsulates the entire enterprise: “these works blur the separation between image and object and exist in a state of flux, they are constantly re-imagined.” He’s right, there is indeed no shortage of imagination to bask in here.

Daniel Cardoso Llach | Miller Gallery at Carnegie Mellon University
Photos: Tom Little & Smokey Dyar, Video walkthrough: Joshua Brown