Off-the-Shelf Genius: What Happens When Parametric Is Ubiquitous?

After spending two days at an architectural-software conference, I expected to leave knowing about a couple of sexy new building projects. Instead, I came away from SmartGeometry, a feverishly technological San Fran­cisco confab, with a new word: parametric. I was invited to attend by Bentley Systems, a Pennsylvania-based software maker. SmartGeometry, however, isn’t the name of one of its products but rather a nonprofit consortium, founded in 2001, made up of representatives from the architecture and engineering firms Foster + Partners, Kohn Pedersen Fox (KPF), Arup, and Buro Happold, plus a number of leading schools. The group and its annual gathering are devoted to building a culture around advanced approaches to computerized design and fabrication. And the main focus of this is parametric software, commercially launched by Bentley in 2007, called GenerativeComponents.

On my first day at SmartGeometry, I sat through six hours of presentations in a hotel ballroom in which young scholars and designers showed buildings, mostly unbuilt. It was architecture as mesh, as neural network, as high-strength knitted fabric. Apparently, everything GenerativeComponents spits out is an undulating surface composed of myriad polygons, suggesting that the entire man-made world will soon resemble the ceiling of the Beijing airport. It wasn’t until a midafternoon coffee break that I cornered someone, a writer for an Australian architecture magazine, and asked, “What does parametric actually mean?” He explained—slowly—that software responds to the information you feed it about the program of a building, its proportions, its materials, and how it needs to perform. You might, for instance, influence a building’s form by inputting data about energy use or seismic conditions. I thought about it for a minute and said, “Oh, you mean parameters?”

Or as Dr. Robert Aish, who developed GenerativeComponents for Bentley and has since become the director of software development for Autodesk, later told me, parametric software calculates the relationships between things. If you change the size of one component, the rest of the building adjusts accordingly. “If I draw a circle with a particular radius and then change the radius, that circle will update. If I made that circle define a roof, the whole roof will update.”

I was reminded of my first encounter with parametric software, back when it went by a different name. In 1998 I interviewed three architects whose computer-driven approach was, at the time, associated with Colum­bia’s paperless studios: Greg Lynn, Michael McInturf, and Douglas Garofalo. They had designed a strangely shaped church for a Korean con­gregation in Queens, and they insisted that they didn’t start out with a particular aesthetic in mind but rather fed data about the building’s requirements into a computer, using software that responded to the data by drawing rounded shapes. At the time, this approach was characterized as “blob” architecture.

But GenerativeComponents isn’t really a direct descendant of the blob. As Aish explains, the early CAD systems were inherently conservative, designed “to help people do what they were already doing with the minimum amount of disruption to the way they thought.” Meaning? The software was good for drawing and positioning ordinary doors, walls, and windows. More adventurous architects and students began appropriating software from other disciplines. In the 1990s, the Columbia studios used animation software—excellent for visual effects, less excellent for structural integrity. Frank Gehry famously borrowed CATIA, a software developed for the aircraft industry. And Aish brought his experience designing software for ships. “They’re very interesting objects,” he notes. “Ships are like buildings, only upside down, and they have lots of rectangular and planar things in them like decks and bulkheads, and they have a very curvy hull. And you have to make the straight and flat bits meet the curvy bits.”

In the late 1980s, working at YRM, a London architecture-and-design firm, with colleagues who later coalesced into the SmartGeometry Group—Lars Hesselgren (now at KPF), Hugh Whitehead (at Foster), and J. Parrish (at Arup Sport)—Aish began to apply ship-building software to Nicholas Grimshaw’s Waterloo station, a sinuously arched rail shed that follows the contours of the tracks. Its design involved a graded series of arches that the architects initially modeled in simple CAD. YRM collaborated with the structural engineers on the project and, as Aish explains it, “helped Grimshaw rationalize the geometry.”

In the real world of 2009, beyond the hothouse of Smart­Geometry, the economic downturn and the credit crisis have hit the architectural profession hard—layoffs abound; even projects in Dubai and China have dried up. Conventional wisdom says that architecture as spectacle, a.k.a. starchitecture, is over. As Blair Kamin, of the Chicago Tribune, wrote in January, “The age of the architectural icon—that extravagant, exuberant, ‘wow’-inducing building on a pedestal—is dead, or more precisely, in its death throes.”

As the techno-savvy backroom crews from firms like Foster and Arup demonstrated just how the wow finds its way into built form, I realized that starchitecture is a complete misnomer. While conceived and marketed by charismatic leaders, so-called starchitecture is actually the product of a small army of brilliant underlings. As readers of this magazine know, the design of a building—especially one that performs all the functions we demand, meets or exceeds LEED standards, and contains a grab bag of uses—requires a collaborative effort among architects, engineers, and fabricators.

“There’s no such thing as a designer any­more,” Martin Doscher, from Morphosis, observed in his presentation. “Design is happening all the way through, until they tighten the last nut.” All the swoopy, curvilinear, geometrically perverse gestures that we associate with the world’s most fam­ous, most audacious architects—those are now embedded in, and enabled by, commercial software: Rem in a box.

By noon of day two, I started reflecting on how we got from the moment in 1952 when the Lever House, New York’s first commercial curtain-walled tower, was completed to the 1970s, when corporate Modernism was the norm. It wasn’t about software back then. All the systems and players—the design skills of architects, the supply stream, the building trades, the real estate demands of developers—synced up to support and perpetuate glass boxes. And now I was witnessing a similar realignment: architectural methodologies and systems conspiring to turn radical ideas into mainstream ones.

I asked Aish a quesion: If older CAD software was conservative (written to enable conventional design), is parametric software inherently liberal (coded to encourage free-form architecture)? Aish wouldn’t go that far, but he acknowledged that the software makes it a lot easier to draw curves and refine them into structures that will stand. When I asked him what happens when the software allows the exotic to become routine, he didn’t have answers, only questions of his own: “If you lower the technical bar to people doing this, do you get the level of thoughtfulness in the design? Before, it was really difficult to do, so only the really cool people thought about it. Now you make things accessible. How do you avoid becoming trivial?”

Aish’s one weapon against what he calls “techno-kitsch” is the SmartGeometry conference itself. He hopes that by grooming the software users, more inventive architecture will result. But as I watch the junior presenters deftly manipulating amorphous buildings to reduce solar gain or in-crease parking, I drift back to the 1990s again and to something John Maeda told me when he began teaching at MIT’s Media Lab about the problem with design software: how users become infatuated with their newfound creative powers. “Designers think, I can make my imagination,” Maeda said. They forget that all the cool things they can do have been enabled by a higher creative power, the software. Or as Maeda put it: “They don’t realize they’re in someone else’s imagination.” What dawns on me at SmartGeometry is that starchitecture is here to stay. Crumpled, folded, rounded show-off “wow” buildings will become increasingly commonplace—but they’ll no longer be designed by the stars. The only true stars will be the guys who design the parametric software.