The Friendly Orange Glow Brian Dear Pantheon: 2017.
“Imagine discovering that a small group of people had invented a fully functioning jet airplane capable of flying long distances at hundreds of miles per hour, decades before the Wright brothers”. So writes Brian Dear in The Friendly Orange Glow, his history of a computer system that most people have never heard of, but perhaps should have. That system, Programmed Logic for Automated Teaching Operations, or PLATO, brought together dreamers, gamers and engineers in a network at the dawn of the 1960s, long pre-dating the Internet. But was this collective venture really as ahead of its time as Dear claims?
The story he tells is both intriguing and a familiar one in the history of technology: a set of determined visionaries break down barriers to make way for a brilliant advance. What differentiates The Friendly Orange Glow is that the vision behind PLATO ultimately failed. The product created was overshadowed, forgotten by all but its most devoted users, and shut down many years later.
As Dear relates, PLATO’s origins go back to an unexpected source: B. F. Skinner. The pioneer of behavioural psychology was famous for his operant conditioning chamber (also known as the Skinner box), in which animals learned to receive food by pushing a lever. He believed that humans, too, would respond to such conditioning, and soon conceived of a ‘teaching machine’ that would allow students to learn through immediate feedback. His 1954 design, a wooden box housing a rotating-disc contraption, allowed users to move through questions at their own pace. It never quite caught on, but it laid the intellectual foundation for ‘teaching’ computers.
The concept got a second lease of life a few years later, when panic over the Soviet Union’s launch of the Sputnik satellite fuelled renewed interest in education and the nascent field of computers. As Dear reveals, in the late 1950s, the Control Systems Laboratory at the University of Illinois at Urbana-Champaign — a military-funded facility eager to emerge from the shadows of covert work — sought to mesh the digital with learning. Scientists there, particularly physicist Chalmers Sherwin and lab head Daniel Alpert, seized on the idea of a “book with feedback”.
In June 1960, the laboratory launched PLATO under the direction of forward-thinking engineer Donald Bitzer, known affectionately as Bitz. One of its key innovations was a graphics terminal: the “friendly orange glow” refers to the colour of its flat-panel gas-plasma display.
PLATO was in some ways inadvertently revolutionary. The initial system relied on ILLIAC, a 5-tonne “formidable beast” of a computer that took up most of a room. Decades before personal computing, it was not feasible to have a classroom filled with computers, so students worked at terminals hooked up to a mainframe.
Thus PLATO was an early demonstration of time-sharing and networking. Yet in Washington DC, Dear notes, something even more intriguing was taking place. A defence organization, the Advanced Research Projects Agency (ARPA), was also working towards connecting computers on a single network. Their eventual system, ARPANET, rejected the mainframe paradigm and linked host computers called interface message processors into a network.
ARPANET and PLATO expanded in parallel in the 1970s — and in isolation. This was, Dear notes, “one of the great tragedies in PLATO’s history”. Incredibly, a PLATO terminal was sitting right next to an ARPANET terminal at the University of Illinois, as the latter network was expanding to universities around the country. Dear compares them to “two televisions permanently tuned to different channels”. But the difference was deeper. PLATO was never meant to communicate with other computer systems; it was just connecting terminals across phone lines. And therein lies the fundamental problem with Dear’s thesis. Unlike ARPANET, PLATO was not designed to be a seed that sprouted into a tree of networked computing. Rather, it was a single branch that grew in the wrong direction, and subsequently withered.
So, even as PLATO attracted growing interest for its capacity to connect users, licensing company Control Data Corporation stuck stubbornly to the original educational goals. And PLATO’s designers remained fixated on ‘dumb terminals’ hooked up to a mainframe computer — the ultimate digital dead end. By contrast, engineers at the influential technology-innovation lab Xerox PARC in Palo Alto, California, believed that the future lay in desktop computers. In hindsight, the rejection of personal computing by PLATO’s proponents is maddening, Dear points out. He quotes William Norris, head of Control Data Corporation, as saying, “We found the proliferations of Apples and IBMs a roadblock to PLATO.”
Although it’s hard to accept Dear’s contention that PLATO was a jet in a Wright-brothers scenario, his prodigious research makes this book a worthy addition to the history of computer science. The narrative suffers from extended quotes at times, but the story shines through — a fascinating tale of missed opportunities and blind spots. PLATO lumbered along, ignoring the “coming freight train known as the microcomputer revolution” that would overtake mainframes and leave graphics terminals choking in the dust.
Dear also rightly bemoans the failure of teaching machines: online education today focuses on less interactive approaches, such as massive open online courses, or MOOCs, which are far from Skinner’s vision of immediate feedback. As Dear writes: “The field of educational technology, largely ignorant of its own history, seems eternally condemned to repeat itself.” Yet when, in 2015, administrators of NovaNET (as PLATO was rechristened) prepared to take the network offline for good, the users who stayed up late to experience its final moments weren’t there for educational reasons. This was an online community of nostalgics. The digital future was always right there in front of them, glowing orange, if only PLATO’s creators had grasped it.
Nature 551, 438-439 (2017)