The Friendly Orange Glow

by Brian Dear

  —

  One of the most successful courseware products CDC developed was a full Basic Skills Learning System curriculum, used in adult training centers, military bases, and other sites around the country. It was probably their most lucrative success in the education market. “One of the major contributions Basic Skills made to PLATO,” says Peter Rizza, who was involved in developing the courseware, “was that it was the first ‘Cyber Classroom’ ever created. What this means is that entire third- through eighth-grade equivalent curriculum lessons were delivered to end users without the need of an instructor. The system ran itself.” And like CERL had discovered, CDC found a market in prisons. PLATO’s Corrections Project that Marty Siegel had led had been deployed in a similar fashion, and on a much larger scale. Siegel found CDC’s Basic Skills instructional design somewhat lacking and its content superficial compared to Basic Skills lessons that his own group produced over the years: “You could sort of go from the beginning of the lessons to the end of the lesson and it would say, ‘Congratulations, you’ve mastered this stuff,’ and you really didn’t. Because it was sort of like, uh, here, I’ll give you a question, um, if you don’t know the answer, I’ll tell you the answer, then you type the right answer in, then you go to the next question. And you could sort of get to the end by not knowing anything, but just sort of copying the right answers in at each time. Our [CERL] lessons had a very different kind of instructional model underlying it, that they were really mastery based, and teaching underlying generalizations. The way it worked was that if you missed an item, that item would come back for review, but you’d get it in a different form. So you could never just memorize a particular item. It would always be a variation on a theme.”

  In time CDC would attempt to re-create the success story they’d achieved with Basic Skills by creating other large online courses for GED certification, and even an entire college-level engineering course called the Lower Division Engineering Curriculum (LDEC). The company also developed a successful series of courses in sales training.

  After much campaigning, Michael Allen was able to convince CDC that it needed to create a line of tools to manage all of this instructional content, and the result was a product called PLATO Learning Management (PLM), a computer-managed instruction tool that today would be recognized as perhaps the world’s first fully fledged learning management system (LMS), a type of software that has turned into a multibillion-dollar industry today.

  “There were very few profitable markets out there because of the extremely high cost of the technology,” says Jim Glish. “CERL tended to think that PLATO didn’t cost anything and everybody could use it because it was cheap. But they had it for free. They didn’t have to underwrite the costs or cost-justify everything they did the way a business did. Control Data was keenly aware of how expensive it was, and was constantly trying to find ways to make it less expensive for the end user, and CERL really wasn’t very concerned about that, I don’t think. I think they looked at it more from the academic purist point of view, of what it potentially could be, but the reality was it had to be cost-effective as well. Control Data was kind of caught between a rock and a hard place, because there were some of the factions within Control Data that kept it more expensive than it needed to be on the one hand, but then there were the market studies that were done and there were really only five key industries that could afford PLATO even up into the mid-1980s.”

  Those industries were aviation, including airlines like American and United; power utilities, including nuclear and electric; the financial industry; manufacturing; and telecommunications. “Basically it turned out that they were the industries that were regulated,” says Glish.

  One successful application of CDC PLATO in the power utility industry began by complete surprise. The evening of Wednesday, July 13, 1977, had started off normally for Luke Kaven and Michael Besosa, who stuck around the Control Data Learning Center on 42nd Street in New York where they worked in order to play some Airfight. Just after 9:27 p.m., the power went out, trapping them in the building. What they did not realize at the time was that the power had gone out all over Manhattan, in one of the largest blackouts in American history, caused by a chain of lightning strikes, transformer overloads, and finally a shutdown at the Ravenswood 3 steam generation plant, sending Manhattan into darkness and forcing La Guardia and JFK airports to close for eight hours. As a result of that disaster, “the state legislature in New York mandated that they had to have all their people trained and it had to be different and better training than they had before,” says CDC marketing manager Ted Martz, “and so we developed some of the best simulations we ever did, some network simulations where a switcher out there would see something going down on the power line and have to make some judgments about a reroute of electrical traffic, and we simulated whether it would work or not work or blow up—very expensive, but it was the only thing on the market that could do what that was doing. And they valued it, to the point of when I think they went out and actually bought their own system.” The simulation was based on a program called the Procedure Logic Simulator written by Luke Kaven, who was in college at the time while working for CDC and had his own IST terminal in his Hampshire College dorm room during the 1978–1979 academic year.

  In the financial industry, training was required to get certification. For instance, one of the most successful uses of PLATO was with the National Association of Securities Dealers (NASD), the company behind NASDAQ. “We did all of the certification testing for them,” says Jim Ghesquiere, who had been in the PSO group at CERL before joining CDC. “In order to be a stockbroker you need to be able to pass a couple of different examinations. And one of the things that NASD wanted to do, of course, was ensure that every time that they gave a test, it was of equivalent difficulty of other tests that they had given, so that I wouldn’t get a test and just by the luck of the draw pull items out of a database and had a very easy test, and when you come, by the luck of the draw you have all hard questions. And you know, one person passes and the other person doesn’t. So they had a large test bank, and when a person came to take the test, we would generate their test on the fly, and do it from a statistical analysis so that we would have pulled questions out of different topic areas, and balance the difficulty, so statistically everyone had exactly the same level of difficulty of the test. We could monitor all that online while they were taking tests, and we kept all the historical data and it was scrutinized a great deal by NASD. But it proved to be, I think, one of the early electronic standardized tests in the country….Probably a couple million exams were delivered on that system.” Merrill Lynch used PLATO for similar purposes. Certification testing turned out to be a lucrative industry for years for CDC.

  In telecommunications, CDC found some market resonance, particularly with AT&T. Another large market, one CDC knew very well going back to the beginning of time, was the United States government. CDC had been selling scientific computers to various government agencies since the 1604—it was one of their primary markets. With PLATO, they had a new way to sell to these agencies as well as the military services, with their constant need to train personnel. “Within Control Data,” says Jim Glish, “there were different marketing groups, so within the commercial marketing group, government and education weren’t allowed to be factored in. That was probably another problem in Control Data…the segmentation of the marketing groups, and the lack of cooperation between them. It meant that some of the resources were duplicated, and some of the strategies were fragmented along the way. And the Government Systems Group had the main five Army PLATO systems, and the FAA, and so on.”

  —

  Things were changing on the hardware front. After the success of the IST, CDC pushed for yet another revision to the terminal design. The result was a smaller, single-unit machine encased in molded white plastic. This terminal, called the IST-2 (which in effect caused the old IST to be forever referred to going forward as the IST-1) had no PC-style expandability lik
e the IST-1, though it did have a microprocessor inside. The move from IST-1 to IST-2 seemed like a message from CDC, a rejection of the coming freight train known as the microcomputer revolution that was bearing down on the entire mainframe industry. “The IST-2 was brought out as a cost-reduced version,” says Jock Hill. “We considered it a step backwards because they put everything on one board, but that was Roseville [the Terminal Services Division], basking in the success of the first terminal, they wrestled for control of the budget away from Bob [Morris], and when they got it, they went back to their old way of doing things, which produced a lot of terminals, at a slightly lower cost, but no technology advance, which was disappointing. Because Bob was looking for advances in technology all the way down the line.”

  CDC would later follow the IST-2 with the similar-looking but more capable IST-3, also marketed as the CD 110, a complete, stand-alone microcomputer workstation with the familiar corporate gray 512 x 512 pixels and touch screen, a gigantic eight-inch external floppy drive, and support for PLATO. Still no support for color graphics. The machine was very expensive, but gave CDC a platform for finally, reluctantly, moving the delivery of PLATO courseware from the mainframe to a micro-workstation, which many customers were clamoring for.

  A common anecdote told among numerous Control Data veterans interviewed for this book involved said veteran walking into an executive’s office, and gently attempting to broach the subject of personal computers—that CDC should be doing something much more aggressively in this area, that it had a chance to own the market if it chose to. These anecdotes always end the same way: said CDC employee being booted out of the executive’s office, amid shouts along the lines of “We are a mainframe company!”

  One example of this attitude was memorialized in a handwritten memo from John Dammeyer about PLATO sent to Norris. After Dammeyer left CDC full-time, he consulted for CDC for a while doing research on the K–12 education market in the state of Minnesota. The result, the memo stated, was “a major proposal to the Minnesota legislature to place a minimum of two hundred terminals in small rural secondary schools for two dollars per terminal hour.”

  The proposal was politically mishandled, and along with the lack of curricular cohesiveness, the proposal fell through; but the study generated a broad awareness of computer-assisted learning. Unfortunately for CDC, the Minnesota schools, and especially Minnesota rural kids, APPLE move[d] in swiftly behind the rejected CDC proposal and, together with school administrators, perpetrated on the kids hundreds of the little APPLE toys.

  The growing impact of such toys appears to have thrown CDC’s PLATO people into a panic to produce their own “stand-alone” version of a functionally cut-down “little PLATO.” This bodes the sacrifice of PLATO’s relative excellence for near future revenues and shares of the educational toy market; and if this is permitted to happen, 20 years of WCN’s [William C. Norris’s] educational foresight and coverage shall have resulted in a tiny fraction of what he dreamed.

  Unfortunately for CDC, these pesky little computer “toys” were not going to go away. If anything, their stampede onto the marketplace announced to all who cared to listen that what was going away were mainframes. In one of the great understatements of the computer era, in 1986 Bill Norris said, “We found the proliferations of Apples and IBMs a roadblock to PLATO.”

  —

  Mark Ciskey was another bright software engineer whom, along with Kevet Duncombe, Chuck Miller had hired right out of Iowa State because, Miller says, “Kevet Duncombe had said, ‘Smart guy, go get him,’ so, sight unseen, I hired him too.”

  Ciskey had an Apple computer at home. “He told us one day, you know, ‘What’s so hot about PLATO, you can do anything you want on this Apple that we can do on PLATO.’ We said, ‘Oh yeah? You can’t play Empire.’ Well, that was on a Friday. Monday he came in and said, ‘Watch this,’ plugged his Apple in, and he had developed a PLATO terminal emulator. Granted, he had to really reduce the character sets and everything [Apple II’s screen resolution was but a fraction of PLATO’s], but it was PLATO. I said, ‘This is incredible.’ Because at the time CDC made its own terminals, at five thousand bucks a pop, and they’re expecting schools to buy one for each student! My folks are both teachers, my sister’s a teacher, and I know that if a teacher gets fifteen spare bucks per year, they’re lucky. They can barely buy chalk for the room, much less buy five-thousand-dollar Control Data specialized terminals. Everybody had Apples, because Jobs was brilliant and gave them away.”

  Miller decided to call a meeting with some thirty CDC marketing managers, sat them down in front of this Apple computer that they were all painfully familiar with, and told them, Watch this. “Brought up the PLATO page,” Miller says, “signed on, went into some courseware, and said, ‘Well, what do you think of that?’ And the head marketing guy—and this is the reason CDC died—sunk his face down in front of the Apple screen and said, ‘That’s green. Don’t like green.’ Quote, unquote. I was stunned, what the hell is wrong with you, and out of that room of thirty people, about three people came up to me and said, ‘You mean I can walk into any school in the country with a disk in my pocket’—or briefcase, ’cause they were big then—‘plug it in, and bring up PLATO on their equipment?’ And I said, ‘Yup.’ The number two person heard that and said, ‘Well, then we’ll have to charge more than a terminal for that disk, ’cause it’s value added. ’Cause that would have to be six thousand dollars for that.’ No, no, no, you idiot, put them in Cheerios boxes, give them away. We don’t want to make money selling hardware disks, we want to make money on connect time. That’s what we are selling! And that idea never sank in.”

  Miller says, “The thing that hurt PLATO the most was the way it was rammed down everybody’s throat as ‘Thou shalt take PLATO and make it prosperous,’ as opposed to ‘Here’s an opportunity, we have to change the way we do business.’ CDC just never got over ‘I want to sell a mainframe for $10 million.’ They never got over that. Their motto was If it plugs in the wall, it’s way too small. Got to be a mainframe, and going to a service was just beyond their comprehension. And they didn’t have any compensation to make them think that way. It was just a mind-set that they were stuck on.”

  Even so, Miller and a band of other like-minded co-workers would not surrender to the corporate mind-set. Instead they took the idea of “PLATO-terminal-on-a-diskette” and their dream of turning PLATO into a home-based online service for plain folks—in essence exactly what Bitzer had been talking about in 1975—and they refined it, polished it, made a version for the IBM PC as well, wrote some documentation, put it in a shiny box, got a few well-placed allies within the company to bless or at least ignore it (explaining to any CDC manager who might raise eyebrows, No worries, nothing to see here, we’re just selling idle mainframe time at night, not going to bother anyone, move along, not the droid you’re looking for) and sold it as a new service called Microlink, eventually renaming the service Homelink. By the time it was squeezed through the corporate marketing apparatus, the pricing had been set to typical Control Data numbers: $45 to purchase the floppy disks and a few pamphlets containing documentation. Add to that an annual fee of $25. Add to that an hourly connect-time fee of $7.75. And if you wanted file space for your own notesfile, or to write some TUTOR code, the fee was 17 cents per lesson part per day. (For comparison, a game like Avatar used somewhere on the order of fifty or more lesson parts.) Bottom line: PLATO Homelink was prohibitively expensive. Miller estimates the subscriber count peaked at only around eight hundred customers nationwide. AOL, Prodigy, and CompuServe, the three big American online services that were about to explode in the late 1980s and early 1990s, never had to worry about Control Data as competition.

  24

  Diaspora

  When Fred Hofstetter came to the University of Delaware in 1973, he was one of only a handful of music scholars in the country who had even dabbled with computers, let alone become fairly proficient at programming. Tall, thin, and bounding with
energy, Hofstetter had first received a BA in music education from St. Joseph’s College in Indiana, and went on to Ohio State University for a master’s and a PhD in music theory. The research for both graduate degrees involved extensive use of computers. For his master’s thesis, he wrote a program that enabled a person to type a musical score into a keypunch machine using a special, machine-readable code. The system then analyzed the music.

  For his 1973 doctoral dissertation he constructed a computerized, musical equivalent of Professor Henry Higgins: given a piece of music, the system could determine what country it was written in by analyzing its stylistic traits. He took sixteen string quartets: four from Germany, four from France, four from Czechoslovakia, and four from Russia. “The dissertation raised a lot of eyebrows,” he says proudly. By the time he finished the dissertation he had left Ohio State after accepting an assistant professor of music theory position at UD, where he rewrote the music-identification program to run on Delaware’s Burroughs computer mainframe.

  News of Hofstetter’s computer projects spread during his first year at Delaware. He soon found himself on an educational computing subcommittee of the faculty senate. “No one ever dreamed what was going to happen, but during that year we started, we had some travel budget, and we traveled around the country, looking at systems. We went to all the main vendors.” The committee’s primary goal was to find an existing, already developed library of courseware materials so they could show people back at Delaware what could be done with computers teaching college students directly. “We looked at the Hewlett-Packard systems. We looked at the Digital systems. We looked at TICCIT.”