Internet Dreams, 1979
THE PHOTOGRAPH AT RIGHT DEPICTS A SCENE THAT AT FIRST glance may not seem remarkable. A woman is sitting in front of a computer terminal; another computer is in the background. Scanning the walls, we see a Boston Red Sox pennant behind her and some drawings next to the terminal. Along with the small, high window, these suggest that she is in a basement rec room. In any case, she is not in an office. The terminal looks old-fashioned and has no mouse or other pointer, but otherwise this could be a scene found at countless homes in the twenty-first century.
The picture was taken in 1979, a critical time in the history of computing. It depicts the moment when computers first became accessible to significant numbers of individuals for personal use at home. In this case, the computer is playing an especially valuable role. The woman in the picture suffers from multiple sclerosis, but she is training in her home in Bloomington, Minnesota, to be a programmer. That explains the instructional manual for COBOL—the standard programming language at the time for business and commercial use—on which her crutch is resting.
A few years before, the illness might have restricted her freedom just as much intellectually as it did physically. Now, however, she can make profitable use of her skills and keep in touch with thousands of other people. In the years since the picture was taken, this notion has become unremarkable, as computers have opened up many avenues for work and play to people with disabilities. But at the time it was a revolutionary idea.
The woman is using a terminal supplied by the Control Data Corporation, of Minneapolis-St. Paul, Minnesota. It is running an elaborate program called PLATO (Programmed Logic for Automated Teaching Operations), which was developed starting in 1959 at the University of Illinois at UrbanaChampaign. At first PLATO’s programmers used the university’s ILLIAC-I computer, a mainframe with what today would be a laughable amount of computing capacity, though at the time it was among the world’s most powerful. Later the project was shifted to a set of very large CDC 6000-series computers designed by the legendary computer architect Seymour Cray, and then to a set of CDC Cyber-series machines. The university, Control Data, and the National Science Foundation were joint sponsors of the PLATO project.
The leader of the effort was Donald Bitzer, who thought PLATO could help solve one of America’s greatest problems: “It became apparent to me and others that the school systems, in the larger innercity schools, were turning out students that were likely to be functionally illiterate in our society. And I was very interested in knowing if it was possible to use our new upcoming computer technology for helping solve this problem.” Training the handicapped and helping them keep in contact with the outside world was not the project’s primary goal, but it fit in with the idea of using computers to help previously excluded people participate in American society.
DOING SO WOULD REQUIRE NOT JUST MORE POWERFUL computers but entirely new software and hardware tools. The PLATO team developed a programming language, called TUTOR, that was especially suited to educational software, and in 1964 Bitzer and Gene Slottow designed the world’s first plasma display terminal for PLATO. The earliest version of the system had used ordinary televisions for display, but they required too much mainframe time to refresh the image continually. The early-1970s version, PLATO IV, had movable, touch-sensitive onscreen objects, like those used in today’s computers except with no mouse. Unfortunately these proved too expensive to manufacture for mass-market use.
Although PLATO was designed with education in mind, games were among its most popular offerings; indeed, the woman in the photograph is examining a list of them. These were an important selling point. Some of PLATO’s games, usually invented by users, were simple interactions between user and computer, sometimes with an educational slant. Others, geared more to adults, could accommodate multiple players at widely scattered terminals. PLATO was particularly good at these applications, since the system allowed for virtually instantaneous response to users’ commands, faster than is available to many Internet users today and far superior to most longdistance computing at the time.
Typically, when playing such games, players would use noms de guerre , fostering the kinds of virtual identities that one finds in Internet chatrooms and game sites today. It did not take long for one PLATO programmer to devise a pure chat program, which he called Talkomatic. Nearly all the social phenomena of current chatrooms— flaming, fake identities, arguments over which topics are off-limits—first appeared in rudimentary form on Talkomatic or on another program, called PLATO Notes. PLATO Notes was originally designed to let users and programmers correspond about bugs in the system, but it soon expanded into a general bulletin board, with annotations and comments in threaded lists. It served as the inspiration for modern “groupware” programs such as Lotus Notes. Yet there was a lot more to PLATO than games and discussion groups. Its mainframes stored a great variety of programs and information, including educational “courseware,” engineering and scientific software, databases, and business applications.
As PLATO’s development progressed, the original partners eventually chose to pursue the commercial and educational ends of the project separately. Control Data obtained the rights to the name PLATO in 1976 and began marketing it in 1978. Meanwhile, the University of Illinois continued work on its own version, which concentrated mostly on education, eventually adopting the name NovaNET. Individual PLATO systems, each with its own operator (usually a university), were scattered around the world, but most of them were in the United States and Canada. Each system was capable of handling up to 1,000 users at a time, and almost all the users were associated with institutions. These individual PLATO systems were linked in a worldwide network, allowing far-flung users to communicate and play games with each other.
PLATO was a bold effort by Control Data to bring computing into a new market where the dominant computer company, IBM, had little presence. Computing at that time was still mostly a matter of batch processing, with programs and data keyed into storage and run through the system by computer operators when the user’s name came up. Many installations still used punched cards. Mainframe manufacturers and software companies supplied remote access to some users through time-sharing terminals, but these were mostly engineering or business customers, not the mass market that PLATO’s marketers envisioned. Control Data knew its computers were among the fastest in the world and could handle ambitious interactive software that its competitors could not. This made PLATO much more promising than the other educational systems then being sold, which relied on installing minicomputers in individual schools.
Like Bitzer, the company’s chairman, William Norris, was a strong advocate of attacking social problems with computers. He stressed the importance of using them in ways that took full advantage of their capabilities. Computer-mediated rote “drill-and-practice” software was already available, but Norris and many others thought this was a waste of the computer’s power. PLATO’s courseware, they expected, would generate a creative and truly educational dialogue with the user, hence the somewhat labored acronym chosen for its name. It would also have the ability to track the performance of individual students and create personalized lesson plans.
IN THE YEARS AFTER THIS PHOTOGRAPH WAS TAKEN, PLATO became a modest success, and at its peak, around 1985, more than 100 systems were in use around the world. To its fans, PLATO was little short of miraculous. Think of how amazed you were when you first started using e-mail and the Internet; then move that reaction back 15 years. As is true with many short-lived technologies, nostalgic enthusiasts insist that PLATO could have been a roaring success and survived to the present day if only it had gotten the proper support. PLATO veterans like to compare its virtually instantaneous response with the often sluggish downloading on today’s Internet. Yet along with its advantages, PLATO had cumbersome features that made it far from ideal, both as an educational tool and as the mediator of what today would be called an online community.
Chief among these was its cost. The University of Amsterdam, for example, spent $200,000 to connect to PLATO in the late 1970s before deciding that the expense and administrative burden were not worth it. One source sums up PLATO’s problems as an educational tool this way: “High cost of courseware development in relation to a course’s limited lifecycle; troublesome user-machine interaction; difficulties of combining CAI [computer-assisted instruction] with traditional instruction methods.” Achieving a critical mass of users would have reduced expenses, of course, but the limit of 1,000 terminals per system (which increased to 4,000 by the mid-1980s), impressive as it seemed at the time, might have interfered with this goal.
The PLATO model started to look outdated in 1986, when the first local-area networks were installed, bypassing the need for a central mainframe. Three years later, with Control Data suffering financial problems—partly as a result of the shift away from mainframes—the company sold PLATO to TRO Learning, Inc., which now markets Web-based learning software under the name PLATO Learning, Inc.
Over the last decade or so, many applications originally developed for PLATO have become commonplace on the Internet, including e-mail, instant messaging, chat, and multiuser games. Courseware and educational games are also widespread, and of course, handicapped people are among the Internet’s most dedicated users. What differs the most from PLATO is that the infrastructure providing these services is almost the exact opposite of the structure of the PLATO system. For one thing, the power of personal computers allows most of the graphics processing to take place inside a user’s PC, not in the mainframe. But what most differentiates the World Wide Web from PLATO is that it is connected not to a central repository of programs and data but to the whole world through the Internet. The Web presumes no central team of programmers, authors, or central databases, no gatekeepers of information; it makes no distinction between content generated by the Library of Congress and something generated by a hacker in a college dormitory room.
The photograph contains an early hint of these changes. In the background sits a personal computer, the Tandy Radio Shack TRS-80. Introduced in 1977, the “Trash-80” was one of the first PCs, but its computing power was minuscule. It had no network connection or communications capability, and its external storage was a modified audiotape recorder, visible to the left of the terminal. The woman in the photo probably used the TRS-80 for word processing, simple games and spreadsheets, and perhaps some programming in BASIC. In the years that followed, however, the power of PCs would grow enormously, to a point where distributed networks of information became cost-effective. The mainframe survived as a repository of data, but the hub-and-spoke model that had been built around it was no longer necessary.
Upon looking back after a decade of mass usage, it is clear that the advantages of the Internet have not come without a cost, as anyone who has encountered dead links, spam, and misinformation can attest. Even Tim Berners-Lee, the inventor of the World Wide Web, has wished for more structure to it, and he has recently been devoting his efforts to taming the beast he helped create. The results remain to be seen, but regardless of how the Web evolves, its nature will reflect the aspirations and expectations shown in this modest photo of a woman learning COBOL on a terminal in her basement.
