February 10, 2005 You There, at the Computer: Pay Attention By KATIE HAFNER http://www.nytimes.com/2005/02/10/technology/circuits/10info.html?position=&; ei=5090&en=7a999c4499970da6&ex=1265691600&partner=techdirt&pagewanted=print& position=
FIRST, a confession. Since starting to write this article two hours ago, I have left my chair only once. But I have not been entirely present, either. Each time I have encountered a thorny sentence construction or a tough transition, I have heard the siren call of distraction. Shouldn't I fiddle with my Netflix queue, perhaps, or click on the weekend weather forecast? And there must be a friend having a birthday who would love to receive an e-card right now. I have checked two e-mail accounts at least a dozen times each, and read eight messages. Only two were relevant to my task, but I responded right away to all of them. My sole act of self-discipline: both instant messaging accounts are turned off. For now. This sorry litany is made only slightly less depressing when I remind myself that I have plenty of company. Humans specialize in distraction, especially when the task at hand requires intellectual heavy lifting. All the usual "Is it lunchtime yet?" inner voices, and external interruptions like incoming phone calls, are alive and well. But in the era of e-mail, instant messaging, Googling, e-commerce and iTunes, potential distractions while seated at a computer are not only ever-present but very enticing. Distracting oneself used to consist of sharpening a half-dozen pencils or lighting a cigarette. Today, there is a universe of diversions to buy, hear, watch and forward, which makes focusing on a task all the more challenging. "It's so hard, because of the incredible possibilities we have that we've never had before, such as the Internet," said John Ratey, an associate professor at Harvard Medical School who specializes in attention problems. Dr. Ratey said that in deference to those who live with clinically diagnosed attention deficit disorder, he calls this phenomenon pseudo-A.D.D. A growing number of computer scientists and psychologists are studying the problem of diminished attention. And some are beginning to work on solutions. Ben Bederson, who builds computer interfaces at the University of Maryland, said his design goal is to generate a minimum of distraction for the user. "We're trying to come up with simple ideas of how computer interfaces get in the way of being able to concentrate," said Dr. Bederson, director of the Human-Computer Interaction Lab at the university. When scrolling up and down a document on a computer screen, for instance, he said, some software causes the page to jump. It's an invitation to distraction, in that it requires the eye to reacquaint itself with the document in order to continue reading. To help people understand the importance of avoiding these kinds of jumpy interactions, Dr. Bederson showed that smooth scrolling was not only easier on the eye, but reduced the number of mistakes people make when, say, reading a document aloud. But some distractions don't need much of an invitation. Take e-mail, for instance. "It's in human nature to wonder whether you've got new mail," said Alon Halevy, a professor of computer science at the University of Washington who specializes in data management systems and artificial intelligence. "I don't think anything else is as compelling to divert attention." Dr. Halevy and others talk about making e-mail intelligent so that it knows when to interrupt the user. "Suppose you trusted your e-mail system enough that you're alerted to an e-mail only if it's really pertinent right now," Dr. Halevy said. "If I knew the right thing was happening with my e-mail, it wouldn't be such a distraction." Dr. Halevy said this is a very difficult problem because it requires sophisticated natural language comprehension on the part of the software. "Completely solving the natural language problem is still decades away," he said, but "extracting useful information out of e-mail is a simpler instance that could make much faster progress." Dr. Halevy is working on what he calls semantic e-mail, which provides some structure to the originating e-mail to make it easier for the software on the recipient's side to understand it and assign a priority. Many people, even the experts, have devised their own stopgap solutions to the attention-span problem. Dr. Bederson tries to read e-mail for only 15 minutes every hour. Dr. Halevy sets milestones for himself and breaks down a large task into small ones. "I say, O.K., I'll finish writing this paragraph, after which I let myself check e-mail, go browse the Web a little bit or make a cappuccino," he said. "If I insert enough resting points between the work, I'm much more motivated to go back to it." Others might say, however, that Dr. Halevy's self-induced interruptions remove him from essential cognitive flow. Dr. Bederson, Dr. Ratey and others often refer to the notion of flow, a concept coined by Mihaly Csikszentmihalyi, (pronounced CHICK-sent-me-hi-ee), professor of psychology at the Claremont Graduate University and the author of "Flow: The Psychology of Optimal Experience" (Perennial, 1991). Flow, in essence, is a state of deep cognitive engagement people achieve when performing an activity that demands a certain level of focus, like writing. Mary Czerwinski, a cognitive psychologist who is a senior researcher at Microsoft, is studying the effect of interruptions on such deep cognitive immersion, with Dr. Bederson. "We're thinking that if you're deeply immersed in a flow state you'll be less amenable to a distraction from an incoming notification, much less likely to even know the notification came through," she said. In related work, other Microsoft researchers are developing software that can learn to gauge where and how a computer user is directing attention, part of what they call the Attentional User Interface project. One piece of software in development learns to assign a level of urgency to incoming e-mail messages while shielding people from messages they can see later - based on an assessment of how busy they are. "We can detect when users are available for communication, or when the user is in a state of flow," said Eric Horvitz, a senior Microsoft researcher who directs the project. For Edward Serotta, as for many other people, the problem is reaching that state of flow to begin with. Mr. Serotta is the director of Centropa (centropa.org), a group based in Vienna that has created a searchable online library of Jewish family photos, linked to oral histories. Part of his job consists of writing lengthy grant proposals, an unwelcome task at best. For the past eight years, Mr. Serotta has used a laptop computer. "That means I can take my ability to dodge serious work everywhere," he said. "I really depend on small technical distractions to keep me away from the things I dread doing." He is currently faced with creating a five-year master plan for his institute at the request of two potential funding sources. The continual checking of his e-mail is rivaled by the micromanagement of his iTunes. "I will certainly do what they ask, but that doesn't necessarily take precedence over figuring out whether I should list Stevie Winwood or Steve Winwood in my iTunes library," he said. Mr. Serotta has four local weather services on his computer's desktop, all of which he watches like a hawk, even on days when he has no intention of leaving his office, which is down the hall from his apartment. "This is vitally important because one of them might be off by half a degree," he said. When Mr. Serotta does manage to find himself in the flow of writing, the stretches of time in which he is focused are what Dr. Czerwinski calls "key cognitive flow moments." Dr. Czerwinski's research group is working to identify the signals that such a moment has ended. "It could be hitting save," she said. "Or it could be the end of a Web search." And this, Dr. Czerwinski said, would be a good time to allow a distraction in, like an e-mail notification. "Most software doesn't take your current cognitive state into account when it lets dialogue through," Dr. Czerwinski said. But such predictive interfaces, as they are called, do not necessarily promise a cure for distraction, even for those more disciplined than Mr. Serotta, as they can be distractions unto themselves that throw the user off intellectual course. "It is the very nature of predictive and adaptive interfaces that the user has to look at whatever the system is proposing and make a decision about whether they want to act on it," Dr. Bederson said. As an example, Dr. Bederson cited word-completion software, like the kind often found on cellphones. "It's a trade-off because you have to look at and evaluate each suggestion from the predictive interface," he said. Dr. Bederson is also skeptical of a predictive interface's ability to know when the best time to interrupt might be. "That's very, very hard for a computer system to guess," he said. Hitting save, for instance, might be the start of a more reflective moment. "And that's the most important time to not interrupt," he said. Dr. Csikszentmihalyi, the flow expert, believes interruptions have their place. "I shouldn't knock distraction completely, because it can be useful," he said. "It can clear the mind and give you a needed break from a very linear kind of thinking." He continued, "E-mail could be a kind of intermittent relief from having to think about things that are not really that enjoyable, but when it becomes a habit so you can't do without it, then it becomes the tail that wags the dog, and it's a problem." Peter S. Hecker, a corporate lawyer in San Francisco, said that when he hears the chiming alert of new e-mail, he forces himself to continue working for 30 seconds before looking at it. Thirty seconds, mind you, not 30 minutes. "Deep thought for a half-hour? Boy, that's hard," Mr. Hecker said. "Does anyone ever really have deep thoughts for half an hour anymore?" You are a subscribed member of the infowarrior list. Visit www.infowarrior.org for list information or to unsubscribe. This message may be redistributed freely in its entirety. Any and all copyrights appearing in list messages are maintained by their respective owners.
