PLATO and CDC: How does a big company full of smart people miss a revolution?
The Friendly Orange Glow: The Untold Story of the PLATO System and the Dawn of Cyberculture (Brian Dear 2017; Pantheon) tells a story of business blindness.
The programmers of the powerful CDC mainframe had all of the technical knowledge, and more, of the PC pioneers, but they didn’t want to drop everything and rush to the PC. The business folks behind the mainframe were similarly mentally locked into their well-trodden paths of sales and applications.
The CDC/PLATO folks actually built a modern distributed system, with a microprocessor in every terminal (“desktop PC”) and communications lines back to a server.
Instead of orange pixels, they were grayish white. The new terminal, called the IST (short for Information Systems Terminal), looked more like an early personal computer. A big, wide, heavy base, with a black grille in front, to which a detached keyboard was connected via a thick cable. On top of the base was a monitor, a special elongated CRT with a square display featuring exactly 512 x 512 black-and-white pixels and, mounted directly over the surface of the CRT’s glass, a reflective, acrylic touch screen with barely visible gold wires crisscrossing across the display. During the nine months of development, the price of CMOS (complementary metal-oxide semiconductor) memory chips had plummeted even further than Hill had anticipated. “According to the really long-haired predictions,” says Hill, “it was going to come down, by six or eight to one, and it came down about ten to one, right when we were doing our development. The result was that we could produce a memory-mapped video terminal, which as far as I know had never been done before, because it was cost-prohibitive.
“We produced what in effect was a PC,” says Hill, “in 1975.” When one considers the year this machine was developed, and compare it to what else was available at that time, it is suddenly apparent that CDC had just leapfrogged over the entire microcomputer field. Here is Hill describing his machine: “[ It had an] 8080 microprocessor, it had plugin cards, it had a separate monitor, with a cable going to the main box, it had a separate keyboard, it had plugin modems, plugin memory, plugin communications, and we even had a plugin disk driver, that wasn’t part of the standard stuff, but we had it networked, so it was revolutionary. And our big problem was producing it at low cost. And we did that. That terminal came in with something like a $ 1,300 cost, in the first few terminals. And that was beyond everybody’s belief.” By the time the IST was ready to be sold to consumers, the marketing people had marked up the price to over $ 8,000, says Hill. It was the beginning of a long line of very bad decisions at CDC. Hill believed the terminal should have been sold for $ 100 above cost. “If we’d done that, we would have flooded the market because people knew they could use it for other things. It would take loadable programs— we could load programs down from the mainframe into that terminal.”
Note the last sentence. The system had the same capability as a modern Web browser that may download a Java or JavaScript program from the server.
The author says that CDC had roughly $1 billion in revenue in 1969 ($7 billion in today’s mini-dollars) so it was about one seventh the size of IBM. Management went all-in on computer-delivered education, which meant trying to sell to governments such as the Soviet Union, Iran, and Venezuela. The U.S. government delayed the Soviet sale due to security concerns and then killed it after the invasion of Afghanistan (imagine how many trillions of dollars we could have saved if we had let the Russians support the secular government in Afghanistan and not supported the Mujahideen!). The Iranian deal fell apart due to political instability:
CERL and CDC created Persian-language support in PLATO as part of the demos, and eventually the Shah’s government agreed to a deal. However, it required that the IST terminals had to be made in Iran (or at least have a decal with “Control Data of Iran” and Persian script on it affixed to the screen bezel). In the end, the Ayatollah Khomeni and the Iranian revolution ended CDC’s hopes in that country. Several of the government ministers, including Prime Minister Amir-Abbas Hoveyda, who had attended the demos back in 1975, were executed. CDC personnel had to evacuate the country, and the company lost a lot of money.
The Venezuela dream didn’t pan out either…
“Venezuela was more corrupt than Iran, if that was possible…. In South America, the Venezuelans were known as the ‘Iranians of South America’ and not just for their oil reserves. You could get anything you wanted in Caracas— anything. Like many CDC international offices, CDCVEN [the acronym CDC used for its Venezuelan business] had its own guy specializing in local bribery and ours was good.” This was CDC’s fixer for Venezuela, “used for more local practical bribery associated with licenses, permits, getting employees and families out of scrapes, etc.” … “My short version,” Smith once explained in an email, “is the PLATO buy became entangled in Venezuelan politics and did not survive the massive political infighting and jockeying for a bite out of it for all concerned (including two or more of our own guys). I do not believe we lost it because we did not bribe. True there was a corporate public effort to clean up our act (I have seen CDC bribe all over the world— even in places like Germany, supposed to be un-bribable) but HQ never backed off of doing business along those lines (anyway it was very difficult to stop the local CDC folks from making deals HQ did not know about). In a lot of countries it was the only way to do business. When the U.S. government started with pressure on U.S. companies to not bribe they started our downfall in the business world….
Are you a big believer in social impact investing? So was the imperial CEO of CDC:
Morris tried to explain to [William] Norris the benefits of pursuing business and education markets at the same time— charging more to business customers so they could charge less to education customers— but Norris did not see it this way. “Norris logically could see it that way,” said Morris. “But his concern was, ‘I’m doing this because I want to make a social impact on education. And if you guys go and turn your attention to selling in the business environment, you’re going to start forgetting about education, and start forgetting about our end goal. I want you to concentrate on education. Okay?’ And so based on that, we did concentrate on education, I still think today if we had sold into the business environment we would have been able to fund more of the stuff that was getting the price down and achieving the educational objectives that we were out to achieve.”
“Addressing society’s major unmet needs” became Norris’s rallying cry, a remarkably progressive mantra for a tech company in the 1970s and 1980s, and one that the rest of the industry and financial world regarded with befuddlement or derision.
In 1984, Randall Rothenberg wrote a profile of Bill Norris and Control Data for Esquire magazine. The article never ran. However, Rothenberg’s recollections of the article’s conclusions shed light on the predicament Norris and CDC were in, particularly with regard to PLATO. “Control Data,” he says, “was an example of what we’d later call industrial policy; its expertise was in seeking government funding for technology projects relating to supercomputing. When the government market for supercomputing for military and economic applications began to dry up (because of, e.g., the advance of minicomputing), CDC, instead of adapting its business model, began to seek new uses within a government welfare structure for its existing supercomputing technology. Using the technology for training, small business development, etc., was a logical extension of this. What CDC could not do was diverge from a model predicated on powerful central control. The whole notion of distributed systems— in computing, in social welfare, in anything else, it seems— was totally foreign to it. So the inapplicability of its technology to the social-welfare aims it was seeking to address was something the company could not work around. Put another way, it had come up with the perfect Great Society solution— twenty years late.”
CDC and PLATO were successful in some markets:
Those industries [that bought and used PLATO] 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.
PLATO was used to train air traffic controllers within the FAA:
The program was so successful— and the FAA’s budget was so tight— that the aging CYBER computer (and the formerly white terminals, yellowing over time under years of fluorescent light) was still running TUTOR lessons twenty years later, well into the 2000s. Some of the original courseware developers would over the years retire only to be rehired by a desperate FAA needing them to fix or update the lessons they’d developed years earlier— since TUTOR programmers had become scarce.
I think the book teaches an important business lesson. Here was a company that had the capital and the people do make near-infinite money if they had simply shown up in the marketplace and given customers what they were asking for. Instead they got fixated on telling customers what they should want and on selling to a particular cluster of customers. They died (company history) during a time period when enterprises all over the world were greatly expanding their IT budgets and when investors were willing to pay absurd multiples of profits for IT vendors.
More: Read The Friendly Orange Glow: The Untold Story of the PLATO System and the Dawn of Cyberculture
Full post, including comments
