Digitally Enhanced | Custom Jobs
The universe operates according to rules—and architecture obeys. The sun comes up; the sun goes down. In July, when the sun is up on the 42nd parallel north, it shines on a building’s facade at a 20-degree tilt, providing about five kilowatt-hours of energy per square meter. If you’re an architect trying to cool your building efficiently, each kilowatt counts; if you’re a piece of modeling software like Revit, you get in the way. That’s what Mario Guttman, of Perkins + Will’s San Francisco office, discovered. He was designing sunshades for some recent projects, and wanted to generate fin shapes straight from solar radiation data, instead of drawing a fin, testing it, analyzing the data, and then drawing it again. But traditional CAD software like Revit doesn’t work that way. It is built for constructing forms. Parametric design, a rule-based approach hashed out in spreadsheets and data sets, doesn’t translate easily into the program’s standard commands. So architects like Guttman build custom work-arounds that modify their software, designing tools to design buildings.
Guttman calls his program WhiteFeet. It connects Revit with the environmental analysis program Ecotect, allowing him to adjust the shape of his solar fins by changing a piece of analytical data, like the time of year or the angle of the facade. “It’s analysis that translates into design rules, that translate into geometry,” he says.
Perkins + Will is just one of many firms coping with the limits of software by building in-house work-arounds. Reiser + Umemoto and Ocean run their conceptual form studies in customized programs. Skidmore, Owings & Merrill’s (SOM) BlackBox Studio and SHoP Construction are special teams within their firms, devoted to modifying the multilayered, interindustry software network called Building Information Modeling, or BIM. Gehry Partners developed an entire software family called Digital Project. “Do people need to know how to customize their programs to be creative? On the individual level, not necessarily,” says Emile Kfouri, Autodesk’s BIM application development manager. “But every firm, the big ones especially, looks at this stuff.”
Like a proud guidance counselor, Kfouri rattles off a long list of success stories, firms that have springboarded off of standard Autodesk software into using custom programs that solve the unique problems of designing and building the world’s most complicated and expensive structures. One firm wrote an add-on that automated seat alignment in a giant stadium, to make sure each spectator had a perfect view. Another rendered the core of a super-tall tower with a click, instead of making designers draw each individual section.
SHoP’s Barclays Center, the future Brooklyn home of the New Jersey Nets, will writhe under an armadillo armor of 12,000 steel plates. “They’re big monsters, and they’re all different,” says SHoP founding principal Christopher Sharples. “We had a digital model, but we had to give each plate construction attributes for our fabricator: information about thickness, bending radii, connection details, tool paths, and weathering. So we wrote a script to automate the process.”
Extracting, analyzing, and sharing data almost always results in a custom job. “Structural analysis on buildings at this level of complexity doesn’t work within these programs,” said Keith Besserud of SOM’s BlackBox Studio, talking about the firm’s warped, stretched-taffy-like tower in Shanghai’s White Magnolia Plaza. “Communicating that information to engineers involves writing scripts to extract nodal coordinates because they can’t be adequately described through conventional dimensioning methods.”
“We love to see it,” Kfouri says of fixes like these. Autodesk knows their limits. In fact, the company has designed its software with customization in mind, building it around easy-to-use application programming interfaces (APIs) that encourage aftermarket tweaks. “An app without a powerful API means you’re stuck with what the developer gives you. On a weekly basis, someone shows me something and I think, Wow, we never thought of that.” It’s digital dog-wagging; in architecture software, innovation comes bottom-up.
It’s always been like this. Architecture software is broadly focused; it’s designed for basic problems, not unique ones. When firms need something special, they’re on their own. As early as AutoCAD 2.18, which came out in 1986, Autodesk bundled its software with a primitive programming interface called AutoLISP—like the early PCs that shipped with a hefty BASIC programming book, Guttman says.
The more general the software’s appeal, the better it sells, of course, but there’s another reason architects often take programming into their own hands. “We’re one of the most impoverished disciplines when it comes to software,” says Brian Lonsway, an associate professor at the Syracuse School of Architecture. Architects’ favorite programs are likely to come secondhand. Rhino was originally developed for jewelers and industrial designers, Maya for animators. “A lot of the customization architects do is pragmatic: taking software that does some incredible thing, but is not quite right for architecture, and just making the damn thing get a set of drawings ready for the contractor,” Lonsway says. “Gehry Technologies [the firm’s software development wing] took CATIA from the shipping, automotive, and aerospace industries. It’s great at making these big curved pieces of steel, but Frank Gehry isn’t Richard Serra. He has to make that into a building.” And so Digital Project was born.
“The shapes were always out there, but until we had the tools to process them, it didn’t matter,” says Dennis Shelden, one of the founders of Gehry Technologies. “Computer programs didn’t create the new architecture, but they allowed it.” The question, then, isn’t how technology changes architecture, but how architects change technology. The retrofitting of programs from other industries, to handle the analytical rigor of parametric design and the complex specifications of contemporary construction, is a symptom, not a cause, of today’s architectural exuberance.
“If you’re using a tool, you can’t just do what it does,” Lonsway says. “Think about a pencil. It’s flexible. You can break it in half, you can sharpen it to a fine point, or use a dull, softer lead. To use a computer without knowing how to program it is like using a pencil without being able to sharpen it. You’re not fully engaging with the tool.” Lonsway teaches programming and design theory to his architecture students at Syracuse—this is only the third year it’s been offered by the School of Architecture. He uses Processing, a visual programming language that’s a bit easier to understand for design students; you don’t have to be a geek to get it, he says. Programming is “in the air” these days—everyone’s hip to it. “There’s a culture of scripting emerging in schools,” SOM’s Besserud says.
Designers might swoon for process, but Autodesk sells results. Publicists there show off their software’s fingerprint on everything from mountain bikes and roller coasters to Avatar’s twirling alien dreadlocks. But behind the digitized sheen, encased within the work itself, are the rules. “These days, if you want to know the name of a restaurant or something, you just tap your iPhone,” Guttman says. “The rules are there, but they’re hidden. Some architects want to just twist or tilt something, and see it happen. But I like the code. It’s a logical expression of the world.”