The Virtues (and Vices) of Virtural Colleagues
For years, space physicist Robert Clauer would trek off to Greenland once or twice a year to gather data on the upper atmosphere. He would fly four or five hours in the back of a cold cargo plane to reach a site where he would sit for days in a trailer crowded with instrument displays. When he wasn’t busy observing, he could step outside to admire the aurora borealis or watch a passing herd of caribou. The experience was rugged, and sometimes exhausting, but it satisfied his soul and his scientific curiosity.
Today Clauer does the same kind of research, but he doesn’t have to go to Greenland to do it. Rather than travel physically, he is now linked via computers in an experimental “electronic collaboratory” project with other space physicists. The electronic links enable them to initiate experiments from their desktops and study data collected from radar instruments in Greenland, Canada, Norway, and the United States and from space-based satellites. Through chat boxes on personal-computer screens, the researchers can put their heads together to interpret data and compare real-time observations to theoretical models generated on supercomputers.
“They are doing something revolutionary in their science,” says Gary Olson, a University of Michigan cognitive scientist who helped develop the collaboratory that the space physicists are using.
A collaboratory, as defined by computer scientist William Wulf, who coined the word in 1989, is a “center without walls” in which users can “perform their research without regard to geographical location-interacting with colleagues, accessing instrumentation, sharing data and computational resources, [and] accessing information in digital libraries.” Underpinning such a setup is computer software-some designed specifically for the collaboratory and some borrowed from other applications-that enables people at various sites to work on experiments simultaneously. Shared access to electronic notebooks and whiteboards, videoconferencing capabilities, and other such technologies enhance the feeling of being “down the hall while across the country,” as James Myers, who leads a collaboratory project in environmental research, puts it.
No one can yet say whether this way of working together is all for the better. Electronic collaboratories do seem to increase opportunities for mentoring; they may raise the frequency and, possibly, the quality of interactions among participants. In some kinds of work, electronic links foster interdisciplinary cooperation, provide access to a wider range of instruments and information, and help narrow the gap between theory and experiments. But something is lost, too, when people interact with instruments and peers from afar. Sitting in front of a computer is no match for traveling to interesting parts of the world, and it’s hard to develop trust in people you’ve met only through electronic channels. What is immediately apparent is that as the future unfolds, ever more people will collaborate electronically.
The first formal discussion of collaboratories occurred at a 1989 National Science Foundation workshop convened by Nobel laureate Joshua Lederberg, president emeritus of Rockefeller University, and Keith Uncapher, senior vice president of the Corporation for National Research Initiatives and dean emeritus of the University of Southern California’s College of Engineering. Today, Olson estimates that two dozen or more collaboratories in science, medicine, business, and education are under way, in fields ranging from diesel engine design to worm genetics.
Through the Michigan-based Upper Atmospheric Research Collaboratory (UARC), for example, scientists share access to radar and other instruments to study “space weather”-phenomena such as magnetic storms that originate on the sun and can interfere with radio and television reception, disrupt electrical-power transmission, and threaten orbiting spacecraft and astronauts. The Collaboratory for Environmental and Molecular Sciences, based at Pacific Northwest National Laboratory in Richland, Wash., allows scientists in different fields and at many sites to work together on environmental problems, sharing analytical instruments, expertise, and a powerful supercomputer. The Department of Energy is sponsoring two new collaboratories that will link researchers at government laboratories and universities. One is aimed at designing diesel engines that produce less pollution; the other will explore ways to make corrosion-free materials. Even school kids are logging on to the collaboratory concept. Northwestern University’s Collaborative Visualization (CoVis) Project gives kindergartners through 12th graders in 11 states tools to explore the earth’s atmosphere and pick the brains of professional scientists.
The collaboratory has important implications for the corporate world as well. Olson says he has talked to “dozens of companies that are interested in using collaborative technologies to support their work.” Finding ways to work effectively with distant colleagues is one of the most important issues today’s managers face, he adds: “All of the auto companies, all of the aerospace companies, all of the big hardware and software vendors, all of the telecommunications companies are worrying about this.”
Users are quickly beginning to appreciate what collabora-tories can offer. Consider how scientists are benefiting from UARC as it evolves. Originally the space physicists just wanted a way to do more experiments without the expense and disruption of packing up and heading to the Sondrestrom Upper Atmospheric Research Facility in Kangerlussuaq, Greenland. That’s what they got. The system’s designers arranged for the Sondrestrom instruments to transmit data over the Internet to the researchers, who were also set up with chat boxes on their screens so they could share their reactions to phenomena as they occurred. And the scientists could send instructions to the crew operating the instruments.
For the first time, the researchers didn’t have to cram their observations into a short visit to Greenland and hope conditions would be right for observations, says Clauer, who’s currently on leave from the University of Michigan to direct the National Science Foundation’s magnetosphere physics program. Then some space physicists began dreaming bigger dreams. Why not bring more radar instruments online? Tap into satellite data? Make it possible to see the results of computer models and observational data side by side in real time?
A new version of UARC tested this wish list during two campaigns in October 1996 and April 1997. About 50 researchers worldwide logged on to sample a space-physics smorgasbord. Users could enter various virtual rooms-one for satellite data, another for radar data and models, others for communicating with instrument operators or getting help with software snafus.
The new, improved UARC also enabled the scientists to couple theory and observations better than before. In the April campaign, for example, the scientists used theoretical models to predict what the atmosphere would be doing an hour later, then compared those with what actually happened. That, says Timothy Killeen, director of the University of Michigan’s space-physics research laboratory, was “extraordinary”-the previous time lag was weeks or months.
Moreover, Killeen expects that space physicists will begin scheduling electronic sessions to analyze data and write up results immediately after observing campaigns, which could move the field ahead rapidly by shortening the gap between collecting data and publishing results. “The best papers that are coming out now put multiple data sets together and do comparisons with theory,” says Killeen. Given all the contacts necessary among the various scientists, the work involved when a collaboratory isn’t used is “a two-year tour de force that exhausts everybody.”
Collaborative technology also offers scientists a way to replay interesting observations or experiments. Toward the end of the ‘97 UARC campaign, a major solar flare sent charged particles toward the earth, creating disturbances in the upper atmosphere-just the kind of thing that makes an atmospheric scientist’s heart flutter. The scientists who started UARC hope to convene an electronic workshop to take another look at the data from that event, doing additional analyses on the data and perhaps drawing in participants who missed it the first time around.
Another intriguing aspect of the ‘97 campaign was that it attracted even more onlookers than participating
scientists. According to Olson, who is also an associate dean of the University of Michigan’s School of Information, many people who checked in occasionally “weren’t active participants. That’s a mode of participation that wouldn’t have been possible in the past”-certainly not in that cramped trailer in Greenland. Now, anyone with even a passing interest in the subject can take a peek behind the scenes of a scientific investigation (as long as an electronic firewall isn’t set up).
That kind of glimpse could spark a kid’s interest in science or help a college student grasp a concept that would be difficult to understand from a classroom lecture. Onlookers from unrelated fields might also offer useful perspectives the space physicists would never think of. Such scenes of broadened communications are playing out in several collaboratories.
For instance, faculty at Northern Michigan University (NMU), a small undergraduate school in Michigan’s Upper Peninsula, are looking to a collaboratory to expand their access to instruments and experts that otherwise would be out of reach. Through a joint project with Cornell University funded by the National Science Foundation, NMU is installing an all-sky camera, an instrument that takes pictures of the sky from horizon to horizon. Scientists scan the images for atmospheric glows-the aurora borealis is an example-that are tip-offs to what’s happening in the upper atmosphere. The school’s location is ideal for such an instrument, but the faculty lack experience in upper-atmospheric research. Through UARC, researchers at other institutions will not only be able to see and study images from the camera, but the NMU faculty and students also will get to “observe other scientists’ research practices, learn the jargon, and get ourselves to the point where we are comfortable asking questions and contributing ideas,” explains William Ralph, a professor emeritus of physics who heads the project with another physicist, David Donovan. “Then, eventually, we hope we’ll be able to do our own experiments, using both our instruments and other instruments that, through UARC, we can operate remotely.” Students who participate will get a chance to “see what good science looks like,” just as they would at a large research university, says Ralph.
And at major research universities, collaboratories can help less advanced students, report Gary Olson and Tom Finholt, an assistant professor of psychology and director of the University of Michigan’s Collaboratory for Research on Electronic Work. They explain that Michigan space-physics graduate students used to do research only in their third or fourth year; with UARC they are jumping into experiments in their first year.
Students and other collaboratory participants are also finding that the setup helps them find mentors in scientists outside their physical institutions. Aaron Ridley, a former student of Bob Clauer, developed such a relationship with Peter Stauning, a senior research scientist at the Danish Meteorological Institute. Ridley, now a postdoctoral researcher at Southwest Research Institute in San Antonio, was an inexperienced grad student in 1993 when UARC was getting under way. During the first experiment he observed, about three weeks after his arrival, Ridley recalls having “no clue to what was going on.” That was OK as long as Clauer was available to answer questions; but when Clauer had to teach a class or take care of other business, Ridley was on his own. He began soliciting advice from Stauning, a constant presence in the collaboratory. Before long, “Peter was kind of like an uncle,” Ridley says, “who shows you the way.”
Mentoring is also a key part of Northwestern’s CoVis project for K-12 students. Faculty and students from the atmospheric sciences department at the University of Illinois at Urbana-Champaign offer advice and expertise to individual students or whole classrooms. As might be expected, some kids don’t seem to appreciate the offer until the night before a project is due. But others strike up long-term friendships, grilling their mentors on everything from cloud colors and snowfall totals to more personal information, such as age and marital status. “They like getting the quick responses-being able to come up with questions and get answers delivered right to them,” says CoVis coordinator and mentor Steven Hall.
And at the other end of the knowledge spectrum, at least one collaboratory has been set up to help experts find assistance in areas unfamiliar to them. The aim of the Collaboratory for Environmental and Molecular Sciences is to help researchers in different fields work together on the cleanup of radioactive and chemical waste at the Hanford site, a contaminated site in southeastern Washington that covers an area more than half the size of Rhode Island. “What we would really like to do is have some mass spectroscopists and modelers and theoretical chemists work with an engineer who can take that knowledge and turn it into a technology and then work with somebody on the site who’s actually going to use it,” says James Myers, project leader for the Collaboratory for Environmental and Molecular Sciences. “If we don’t happen to have an expert in a certain area, we can go to a university and get one and plug [that person] into the team. It allows us to think more broadly in terms of what we can propose to do.”
Bytes Equal More Ideas
Evidence for the communication power of electronic collaboratories comes not just from users but also from behavioral scientists such as Finholt and Olson who are peeking over collaborators’ shoulders, eavesdropping on their electronic and face-to-face conversations, and analyzing everything from publication patterns to the development of mentoring relationships. They share their observations with programmers who can then integrate changes into the technology underlying collaboratories.
One of the first questions Finholt and Olson asked the programmers was whether collaboratory users talk about the same things in online conversations as they do face to face. They do. “On the whole, when something interesting in the science happens, that’s what dominates the conversation, just as it does in face-to-face conversations,” says Olson. “And similarly, when things are quiet, you find some socializing, talking about schedules, the same kinds of things [scientists] talk about face to face when nothing is happening.” Conversations about families, books, news events, and the like are as common over the electronic channels as across the lab bench.
Electronic collaborators are more likely to share ideas, however. “When groups use computer-mediated communication in brainstorming tasks, they outperform face-to-face groups in terms of the number of ideas generated and, according to some studies, the quality of ideas,” Finholt says. He’s not sure why that occurs. Perhaps, he speculates, “it has something to do with seeing the ideas on a screen-that they’re more visceral or more real.”
Or the reason could lie in the fact that collaboratory participants take care to compose their thoughts and think their ideas through before broadcasting them, suggests Olson. Face to face, people are more likely to blurt out short statements, sometimes without carefully considering what they’re about to say. Although people sometimes send hasty, rash messages by e-mail-as any user of the medium can attest-research shows that such comments are actually less common than in face-to-face encounters. In an electronic exchange, the average contribution is longer, more complex, and more carefully developed.
Another possibility is that the usual cues that communicate status are absent and everyone feels free to contribute. Such status cues do seem to interfere with face-to-face interactions among medical specialists, one of Finholt’s students, Stephane Cote, has found. According to Cote, face-to-face communication between radiologists and clinicians is often hampered by differences in “identity functions”-the mores and ideals that lead to a sense of professional identity.
“The clinician is primarily oriented toward curing the patient,” explains Finholt. “The radiologist has that as a high-level abstract goal, but is mainly interested in rendering the most precise and accurate interpretation of
a particular image. Conflict arises when the radiologist is reluctant to speculate beyond the bounds of what he or she sees, to help the clinician determine what the course of treatment ought to be. To the clinician, it looks like the radiologist is stonewalling or somehow not coming clean.” To add to the problem, these discussions usually take place on the radiologist’s turf, where the clinician is an outsider. Rather than focusing on what they’re discussing, the two specialists often try to reinforce their importance by making cutting remarks about each other’s areas of ignorance, Cote has observed.
Cote and Finholt believe a medical collaboratory recently begun at their school may lessen such tension. Looking at x-ray and ultrasound images on a computer screen, a radiologist can use an on-screen pointer and record oral notes to indicate areas of interest. The clinician can later call up the images on a computer in a clinic and replay the radiologist’s remarks, following the pointer to see the exact area being described. The two physicians can exchange comments, questions, and clarifications as often as necessary without venturing into each other’s physical territory.
Although the system is too new for Finholt and Cote to draw conclusions about its use, they suspect it will help the different parties focus on solving patients’ problems. “It would be an interesting paradox,” says Finholt, if “the elimination of face-to-face communication helped them talk better.”
E-Mail doesn’t Equal Trust
Despite such attractions, behavioral scientists and others recognize that electronic collaboratories have some downsides. For instance, even the best collaborative technologies seem to be poor substitutes for a handshake and eye contact. Eleana Rocco, a visiting researcher at the University of Michigan from the University of Venice, is comparing the formation of trust in groups that communicate only electronically with ones that have met in person. Rocco has had groups of subjects play an electronic game that involves both cooperation and competition. Some groups met face-to-face for 5 to 10 minutes before playing the game; others met only virtually. The groups that had face-to-face contact showed much more cooperation than the others.
“In this kind of task, cooperation arises from confidence and trust in your colleagues,” says Gary Olson. “It’s clear from this research that [trust] requires face-to-face contact. It doesn’t have to be a lot, but groups who only have electronic contact never form trust in the same way that groups who have some face-to-face time do.”
Ford Motor Company learned that lesson the hard way. As part of its plan to reorganize the company by functional rather than geographic groups, Ford tried to use collaborative technology to build international teams. Olson’s group studied one team that communicated almost exclusively through videoconferencing, e-mail, and other electronic means. “After a year, that group had never really become a group,” he says. “It was clear that as time went on, the characteristics of communication among those in the same physical location were quite different from [the group] with people at remote sites.”
Now the company has taken another approach, Olson says. While still using collaborative technologies, it has teams work together initially. That seems to be helping. “There certainly are teams that have face-to-face experience and still have trouble-it’s not a guarantee of success,” points out Olson, “but it does seem to help with the process of team formation.”
A shared sense of trust and identity could explain why space physicists have embraced the collaboratory concept while researchers in less cohesive fields have been slower to come on board. With a long tradition of sharing instruments and data and a clear consensus on such matters as ownership of research results, the small group of space physicists who used the original version of UARC was strong on international collaboration from the start. They had worked together to develop “rules of the road” and had put those into writing. They simply had to modify these rules slightly to fit the electronic forum. But when Olson met with a group of neurophysiologists, geneticists, epidemiologists, and clinicians who wanted to form a collaboratory to study mood disorders, he found a different world.
“These fields all have very different traditions about who owns data, who gets rights to data, and so on,” he explains. Even within each discipline, the traditions were established informally, not by explicit consensus. Before they can begin to collaborate, these groups will have to agree on ground rules-a process that will no doubt require face-to-face meetings and could take considerable time and effort.
Another potential problem with electronic collaboratories is that high-powered researchers and engineers may become so swamped with requests that they flee from this mode of communication and so turn collaboratories into a sort of scientific ghetto. In a January 1997 paper in the journal Psychological Science, Finholt and Olson cite research showing that elite scientists using forms of electronic communication such as e-mail are more likely than their nonelite counterparts to receive messages-but also more likely to ignore them, especially when they come from nonelite researchers.
Of course, the busier the researcher, the more apt he or she is to receive messages of any kind. And scientific snobbery existed long before collaboratories came on the scene. The question is whether the technology makes the problem better or worse. While the answer isn’t in yet, Olson suggests that elitism isn’t the only factor to consider. Some driven researchers may keep their distance from collaboratories because they will not want to take time out to learn how to use the new technology. In competitive fields such as high-energy physics and molecular biology, “there are Nobel Prizes at stake,” he says. “Anything that would slow you down or interrupt your work is a real big risk.” That, in turn, could make more junior scientists wonder, “Should I waste my time learning all this stuff while they’re off earning their Nobel Prizes, or should I get cracking in the laboratory?”
Olson believes that they can do both. Because most younger researchers are more comfortable using computers, they don’t have to invest as much time learning to use collaborative technology as the most notable and generally older people in a field do. The younger workers should quickly find that the technology can speed rather than impede their progress.
But this raises another question: Could students and junior researchers end up relying too heavily on electronic collaboratories? Basking in the glow of a computer screen is not the same thing as experiencing an aurora borealis firsthand. Robert Clauer laments that his former student Aaron Ridley completed his graduate degree without ever making the trek to Greenland in the back of a cold, dark cargo plane or stepping outside the trailer to see the aurora. Clauer notes that collaboratories may make for good science, “but it’s better for the soul to be there.” And on a practical level, the arduous experience has long been an important part of a young researcher’s education in understanding where data come from and what’s involved in keeping instruments running.
Ridley admits he missed out a little. His research experience, he says, was “like driving a car in a video game, versus driving a real car.”
Neither newly minted scientists and engineers nor graybeards are likely to see electronic collaboratories as a real alternative to traditional ways of working unless they can be confident that the underlying technology is reliable. Even collaboratory enthusiasts admit that, with collaboratories still in experimental stages, that isn’t always so. Screen displays change; tools such as sticky notes and other annotation doodads are added and then taken away if they prove unwieldy or unpopular. Such moves unsettle people accustomed to using instruments that look and perform the same way day after day.
Much of the problem is not with the collaborative software but with the Internet. For instance, Olson admits that Internet congestion has seriously interfered with UARC’s performance. But the situation could improve dramatically with the coming of Internet2, a high-speed computer network dedicated to research and education applications that more than 100 universities are building.
How will electronic collaborations fit into the future of the laboratory, classroom, and workplace? The answer will depend on the creativity of those who design and use them. No one is suggesting that the new approach should completely replace traditional ways of working together. Meetings, workshops-even a limb-numbing plane ride to Greenland and the awesome sight of an aurora-will continue to have their place. Just as scientists often must juggle variables to figure out models that best describe what’s happening, collaborating groups must keep tweaking the equation to find the right balance of face-to-face discussions, hands-on work, and electronic communication. The results should be new ways of working that raise productivity, foster creativity, break down barriers while building trust-and still manage to satisfy the soul.
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
Deep learning pioneer Geoffrey Hinton has quit Google
Hinton will be speaking at EmTech Digital on Wednesday.
Video: Geoffrey Hinton talks about the “existential threat” of AI
Watch Hinton speak with Will Douglas Heaven, MIT Technology Review’s senior editor for AI, at EmTech Digital.
Doctors have performed brain surgery on a fetus in one of the first operations of its kind
A baby girl who developed a life-threatening brain condition was successfully treated before she was born—and is now a healthy seven-week-old.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.