Athena didn't revolutionize MIT, but it's here to stay

$100m project still short of promise but for many it's indispensable

Richard Saltus
The Boston Globe

October 28, 1991

In the early 1980s, when the potential of the "computer revolution" seemed almost limitless, the Massachusetts Institute of Technology asked a $100 million question.

What would happen if the institute installed a campus-wide network of powerful desktop computers, giving students and professors easy access to computerized number-crunching, word-processing, instructional programs and electronic conferencing?

It was trumpeted as a grand experiment, a bold initiative to test the premise that computers could help students learn better and faster. Optimists foresaw "sweeping changes" in the teaching of everything from engineering to astronomy to English.

"This may be the largest step forward in MIT's long history of contributions to education," declared then-president Paul Gray in 1983 as he launched Project Athena amid great fanfare.

Eight years later, the experimental phase of Athena has concluded, and the system is at MIT to stay. Access to a network of 800 computer workstations scattered in clusters across the campus is a fact of life for most MIT students. For many, it is an indispensable resource.

Compared with the original grand vision, however, almost everyone agrees Athena thus far has failed to meet its potential; stories of radical change are few.

Several showcase instructional programs have made their mark, however, in engineering and physics and -- perhaps surprisingly -- writing and foreign languages. But other departments, such as the school of science, have found little use for it.

Of MIT's 1,500 course offerings, only "about 200 or so have been touched by Athena, and perhaps 10 were touched in a deep and successful way," said Gregory Jackson, who led a study of the experiment last year.

Few teaching programs have emerged, in large part because planners failed to anticipate the enormous amount of time and effort it would take professors to design instructional software -- by one estimate, 100 to 200 hours for each hour of use.

"The thing we underestimated was the time and cost of building good educational materials," said Steven Lerman, former Athena director, who is now seeking more efficient and less costly ways to devise software. "Perhaps there was an element of naivete" in thinking faculty could create complicated programs in their spare time.

Faculty were reluctant to expend a lot of time on work that likely wouldn't lead to tenure or promotion. Writing software counts neither as teaching nor publishing, said MIT sociologist Sherry Turkle in a 1988 report co-authored with Prof. Donald Schon. Some faculty also complained it was impossible to stay on top of their research if they were deeply involved with Athena.

Even at a technology-minded school like MIT, many users found Athena intimidating, irrelevant, even hostile. With its heavily centralized structure and placement of computer clusters in public areas, Athena gave short shrift to small, independent activities, said Turkle.

Some drawbacks have been corrected or are being changed. Students say the workstations are now far more user-friendly, and the system as a whole is more reliable.

Beyond question, the MIT investment has paid off in many ways; if Athena has not been a panacea, it has opened an era whose dimensions remain uncharted.

Athena was launched shortly after Carnegie-Mellon University began installing a similar network known as Andrew. Said William Arms, Carnegie-Mellon's vice president for academic services:

"A lot of people will say Athena or Andrew didn't live up to their expectations; we haven't reduced our curriculum to three years instead of four, for instance."

Given that it's taken nearly a decade just to build the networks -- longer than expected, "I think both projects have been tremendously successful," said Arms. Students and faculty have access to hundreds of library catalogues across the country; electronic mail and bulletin boards; word-processing and "lots of computing power" for mathematical problems. "Add it up and it's had tremendous impact," he added.

Project Athena was conceived at the time Carnegie-Mellon, Brown University and others were planning to put computers into the hands of students in a big way.

Until then, students had used terminals connected to a central computer, a system known as time-sharing. Athena and Andrew are based on "distributed processing," meaning that independent but linked computers are widely available.

Logging onto an Athena computer gives a student access to a "locker" containing personal files, the capability to perform a wide variety of operations and to communicate with other users or tap data banks -- some of them far from MIT.

Some Athena workstations are equipped with multimedia displays allowing users to watch video or laser disks images, work with interactive graphics and carry out complex computing on the same machine.

Workstation clusters were installed in heavily frequented buildings like the Student Center, which also contains shops, lounges and recreational areas; others were set up in department offices, and a few in dormitories as an experiment.

The $100 million cost was substantially underwritten by Digital Equipment Corp. and IBM, but MIT named its own team, headed by Lerman, to develop and run the program "so that we could control our own destiny," he said.

There is no question that Athena achieved a broad presence on campus; the bigger issue is how deep its impact has been.

In the fall of 1990, MIT officials say, 3,500 users accessed Athena daily. It is used by nearly all 4,389 undergraduates, two-thirds of the 5,239 graduate students, and a third of the 941 faculty, they add.

And as with computers in many other settings, linking the entire campus on a single network has introduced some wrinkles into the MIT culture, among them a new style of conversation and communication. So convenient is electronic or "E-mail," that many people have become addicted to computer chatting -- with friends as close as the next Athena cluster or at schools and businesses across the country.

"I usually drop into a cluster several times a week to check on my E-Mail," said Jabin Bell, 21, a first-year graduate student, who uses it to keep in touch with friends on campus and off, and with his girlfriend at Wellesley College.

And like computers everywhere, Athena gets used for games. "At night after studying so much that my brain has gone dead," said Bell, an engineering student, he turns on games like Tetris, which involves fitting together geometric shapes.

The network gives professors an easy way to contact students about coursework, cancelled classes or other matters. Students can also consult tutors electronically if they're having problems with a course or reach a professor directly whenever he's on the system, rather than legging it across campus during the narrow window of office hours.

"It enables a student to ask a teacher a question without worrying what his face looks like," said Jackson, who is director of MIT educational studies and special projects. "And you hear what the prof is saying without having to see his `God, what a stupid question' expression.' "

Surprisingly for a school so technically oriented, Athena failed to win the hearts and minds of a number of faculty. Some were soured because of delays in hardware delivery and constant changes and system problems in the early years.

But sociologist Turkle also notes that some are ambivalent about computers as a whole, having seen students become technologically obsessed "hackers." Some professors, according to a book on Athena published last year by Digital, worried that computers would detract from teaching "because students would confuse mathematical models with reality and become further removed from the physical world."

Turkle said Athena's centralized structure initially prevented large segments of the MIT community from linking up.

For one thing, Athena was designed to unify computing and electronic communication. That meant restricting it to certain computer languages, chiefly UNIX, a research-oriented language. But the personal computers built by IBM and Apple, which were just catching on when Athena began, use DOS, a language not suited to Athena, and so personal computers could not easily link up. As a result, many faculty and students brought along their own PCs or MacIntoshes.

"The failure to achieve a coherent educational network reflects the existence of diverse computer cultures at MIT -- a diversity Project Athena tried unsuccessfully to fight," Turkle wrote.

Even so, say both Turkle and Jackson, the head of the Athena study, the project has produced some exciting teaching devices.

In physics and engineering courses, professors designed programs that allowed students to plug in data gathered in lab experiments and have the computer simulate scientific principles as if it were the "real world."

"The computer allowed students to run through many variations of the same experiment," said Turkle's report. Aerodynamics students, for instance, could design airplane wings and "fly" them on the computer, making them do maneuvers impossible in a wind tunnel, just to see what would happen.

One highly successful program, called "Growltiger," allowed civil engineering students to build visual representations of structures such as bridges and observe what happened when loads were applied to them. The program saved hours of calculations, inspired them to try novel ideas, and, unlike a pencil-and-paper exercise, enabled them "to make many small changes in proposed structures and quickly see their effects," Turkle said.

Turkle said the lesson of Athena is that colleges need to produce a flock of "computer-fluent" teachers, an entirely new kind of "educational professional."

For Athena and for similar networks, the future is difficult to predict, but those who have been deeply involved, like Arms at Carnegie Mellon, believe the experiment is not over; it is just beginning, now that the systems are in place and running well.

"I think we're seeing a nice second wave" of more modest, cheaper instructional software, he said. "I am absolutely convinced that {computerized education} is going to have profound implications.

"And, I'm also convinced it's going to take longer than we thought it would."

Copyright 1991