June 1, 2007
Dreaming in Code
Jonathan Harris distills the Web’s infinite avalanche of thoughts, facts, and feelings into exquisitely framed portraits of humanity.
One late-winter morning Jonathan Harris sat in front of a MacBook in his Brooklyn apartment and opened the Web site We Feel Fine. Candy-colored digital balls and dots bounced around a black background. He clicked on one, and it shattered into a spinning sentence of text that flew to the top of the screen: “I feel for you.” Then he clicked on another, and another—all statements beginning with “I feel,” harvested live from the Web’s millions of blogs by this artwork he cocreated:
I feel so lonely today.
I feel he is there watching over me and even more so when I am in the garden.
I feel like I’m in tenth grade all over again.
I feel your might; I only have relinquish’d one delight to live beneath your more habitual sway.
That last one is from William Wordsworth’s “Intimations of Immortality.” Janiele, a 23-year-old in Radford, Virginia, posted it on her MySpace blog “twenty-one minutes ago.” Unsurprised by the serendipitous beauty his Web site creates, Harris then clicked on “Montage,” and a grid of colored squares dissolved one by one into thumbnail images, the visual confetti of the blogosphere: American Idol contestants, the Eiffel Tower, prom pictures, a puppy. Clicking on one created an elegant little explosion of more candy-colored dots, as the image—a corgi frolicking in the grass, suddenly worthy of Tibor Kalman—expanded to fill the whole window. Clicking again launched the page the dog came from, with an intimacy so sudden it felt like magic: “Bron’s Blog: Knitting Without a Net.”
Then the wizard pulled back the curtain. “We realized there was all this amazing humanity hiding on the Web, but most people considered it to be a cold, inhuman space,” Harris explains, speaking for himself and his frequent collaborator, Sepandar Kamvar, whose day job is technical lead of personalization at Google. “So we asked, ‘How can we systematically quantify feelings using the Web?’” In a process Kamvar describes as “not quite rocket science,” they wrote a program that scrapes new blog posts, looking for the statement “I feel.” With the duplicates thrown out, it yielded 20,000 “feelings” a day. Harris punched up the raw feed of XML code in a source file—in other words, plain text. “Just in this form we could tell it was amazing material,” he says. “Which is when the next big layer comes in: how to visualize all this information.”
We Feel Fine—together with the handful of Harris’s other works—defines a profound new kind of information design: it whittles down the world’s 70 million Web sites and blogs into a framed image of humanity. And it does it live, continuously, and autonomously. Architects and designers have experimented with computational design, letting a computer run through a spectrum of possibilities within a given set of parameters. But Harris’s creations are different: rather than static buildings, magazine covers, or shopping bags, they are constantly changing artistic responses to a constantly changing world. By using the Web as both site and material, they offer a way of seeing rather than merely being a sight.
If you believe that the Internet is a cultural revolution on the level of modern capitalism, the nuclear age, or even the age of reason, then think of Harris as struggling to create its Impressionism, its Abstract Expressionism, or its neoclassicism—struggling, in other words, to develop a new artistic language for a new human condition. And undoubtedly for a new generation. At 27 Harris is different from those of us even just a few years older who made it through high school without e-mail, college without IMs, and at least a few years of our twenties without blogs. The material of experience has changed. The old rituals of memory—photographs, scrapbooks, diaries, letters—have moved onto the Web, opening them up for a new kind of analysis. “The goal for me is really to hold up a mirror to the world, and then open that mirror up to the largest number of people possible,” he says.
As an undergraduate at Princeton, Harris was consumed by traditional visual arts. He painted oils and watercolors, and obsessively kept a sketchbook. But he found the physicality of the medium literally limiting—“limited to what my eyes could see, and limited to the people that I came in physical contact with,” Harris explains. So he turned to computer science and began experimenting with its techniques to represent the world. “I wanted to get away from the heavy hand of the designer or painter projecting a single viewpoint, and move toward a model where I was basically creating the tools and platforms and systems that could be interpreted and used in many different ways, based on the viewer’s inputs,” he says. He aligned himself with the mantra of computer artist and composer Golan Levin: to design tools that are “instantly knowable but infinitely masterable”—like the pencil or the piano.
Harris first garnered international attention in 2004, with the Web site 10×10, which he created while a fellow at Fabrica, Benetton’s technology-based art center housed in a Tadao Ando–designed villa near Venice. Every hour 10×10 scans the feeds of major news sources from around the world. Then, in what Harris describes as an “elaborate process of weighted linguistic analysis,” it determines the hour’s 100 most important words, matches each word to an image, and arrays the images in a ten-by-ten grid—all set cleanly against a white backdrop with a Helvetica title. Mousing over each image makes the word that prompted it fluidly expand along one edge—like a finely woven fishing net for the world’s bad news: kill, Afghan, Bush, raid, troops, suicide.
Last fall Harris created a special version of 10×10 for a group show of work from Fabrica at the Centre Pompidou, in Paris. It split the world into three panes, each harvesting a different set of news feeds: West, East, and Middle East. The morning the show opened, North Korea detonated a test nuclear device, and the East and West lit up completely with images of Kim Jong Il and the Taepodong-2 missile site. There was no other news; but the Middle East was silent. “You have to have a system waiting to observe and reflect,” Harris says. The things he makes are “ready for something like that to occur, but they don’t know when it will occur. But when it does occur they’ll be sure to capture it.”
And crucially, Harris made 10×10, but then he set it free—leaving it alone to reveal those moments and make them profound. “It gives you a temperature of the world,” says Museum of Modern Art design curator Paola Antonelli, who plans to include Harris in Design and the Elastic Mind, an exhibition opening at MoMA in February that will explore the intersection of design and science.
Since going live, 10×10 has received more than ten million visitors. One of them was Upendra Shardanand, an entrepreneur who made his name (and fortune) with personalization software he developed as an undergraduate at MIT and sold to Microsoft for a reported $40 million. (It later became MSN Passport.) As a self-described news junkie, Shardanand had been watching media evolve on the Web—and quickly go stale. “As an experience, news was due for a broad change,” he says—something that rendered it endemic to the medium. He e-mailed Harris out of the blue. “Obviously, Jonathan was working on the same problems,” Shardanand says. “And he intuitively understands the Web as a medium—like a sculptor who understands the properties of clay.” In early 2005 Shardanand founded Daylife with a substantial investment from the New York Times, Harris as design director, and a shared insight on a Google scale: to reflect reality, even as reality changed around it. At heart Daylife would be an R&D shop for media. In practice it would sell a back-end service to newspapers and online publications, and run its own news Web site.
Shardanand is cagey about how Daylife works—and the New York Times remains silent about the nature of its involvement, except to say that it’s an investor. But as it exists now, Daylife looks like a high-bandwidth version of Google News. Headlines are laid over large images (as in 10×10 and We Feel Fine); and when you click on a story, the text is joined in an adjacent column by pull-quotes from the story and thumbnail images of “connections.” When sold as a service to publications, Daylife software (operating in the background) automatically draws connections between stories and allows readers to navigate stories over time. It’s a compelling digital enhancement to the adjacencies inherent to print newspapers—which until now have only been weakly hinted at online with “related stories” lists.
But Harris’s ambition for it is unapologetically grand, even boundless. “My vision for Daylife has always been that you could create a system that would be the definitive resource on everything in the world,” he says. Google wants “to organize the world’s information and make it universally accessible and useful”—and that no longer seems like such a stretch. Harris’s mission is, in many ways, rooted in Google’s limitations. “They present their system through words. The word is the lens. It always comes back to the word. You can’t have the lens at seventy-three degrees Fahrenheit, or the lens of despair and sadness,” he says. “There’s all these other kinds of concepts and realities that exist in the world that should be usable as search queries, but right now we’re limited to keywords. And that doesn’t do justice to reality. It’s a good start, but I don’t think it’s the endgame.” Harris points to a small potted plant on the windowsill in his apartment. “It would be nice to be able to rip off a leaf, and use that leaf as a search query to find out about the leaf and everything that matters to the leaf—so you’ll probably get some information about sunlight or soil moisture, or all these things.”
In March, with Shardanand’s blessing, Harris took Daylife’s news-processing power and re-thought its output on his own terms. The result is Universe, a cryptic and elusive response to Harris’s big questions. Against a dark-blue, milky background that represents the night sky, Universe creates starlike “constellations” out of keywords, people, places, and ideas in the news—Bono, Iraq, climate change. You choose one for the center, and the related topics circle around it until you choose another and the orbit shifts again. If Google wants information to be free and accessible, Harris wants it all to be connected—like the way the world really is. He describes this as “infinite monarchies of information,” where each “morsel” has the capability of being “king,” existing in relation to everything else. “This notion of a static rock of a page just doesn’t apply,” he says. “I think a much better approach is to have a system where any item can be the center, and everything else can relate to it.” Daylife does this with news articles, and Universe does it with its flexible “solar systems” of topics (the metaphor is blunt). But it’s delicate—maybe too delicate, too scattered to be powerful. Unlike the raw emotion captured by We Feel Fine, or the amazing distillation of 10×10, the news feels small in this context—disposable rather than profound. Maybe it is. In May, Harris began a five-month sabbatical from Daylife. “I still believe in Daylife. I just think it’s going to take some work to get it there, and there are other things I want to do,” he explained.
Yet it’s not clear the Web can do all this—that it can be a mirror of reality. Not yet at least. Its contents remain uneven and incomplete; not ev-erything in the world exists online. But pointing that out—which you’d think would be obvious—I worry that I’m not grasping what has become self-evident to every 17-year-old: the Web isn’t a thing unto itself but a whole new reality, awaiting its proper representation.
There’s a short story by Jorge Luis Borges that anyone who obsesses over maps will know: “On Exactitude in Science.” It imagines an empire so obsessed with cartography that it made a map of the empire on the same scale as the empire. “The vast map”—of course—“was useless.” It’s a cautionary tale on the limits of representation and, I’ve always thought, an affirmation of the beauty of reality. But what if that caution doesn’t hold in the digital world—where no relations are fixed and scale is not itself a limiting factor (70 million blogs, 20,000 feelings…)? If that’s the case, then the act of representation, of information design, is forever changed. I put the question to Harris. “Trying to depict everything is a fool’s game, and ultimately not that interesting—because it’s just as confusing and complicated as life. So then the task becomes limiting your scope, and within a limited scope providing amazing complexity and depth. That’s this process of ‘lens making’: coming up with a lens that you can point at all of reality but that only lets through certain things. That process is digital storytelling. It’s a process of exclusion—not a process of mimicry.”