The computer scientist Bret Victor gave a keynote back in 2013 that I return to again and again. (See? Keynotes need not be a waste of time and energy!) In “The Future of Programming,” he offers a history of programming – or more accurately, a history of programming developments that were never widely adopted. That is to say, not the future of programming.
The conceit of Victor’s talk: he delivers it as if it’s 1973, using an overhead projector in lieu of PowerPoint slides, and the future he repeatedly points to is our present-day. With hindsight, we know that the computer languages and frameworks he talks about haven’t been embraced, that this future hasn’t come to pass. But as Victor repeats again and again, it would be such a shame if the inventions he recounts were ignored; it would be a shame if in forty years, we were still coding in procedures in text files in a sequential programming model, for example. “That would suggest we didn’t learn anything from this really fertile period in computer science. So that would kind of be a tragedy. Even more of a tragedy than these ideas not being used would be if these ideas were forgotten.” But the biggest tragedy, says Victor, would be if people forgot that you could have new ideas and different ideas about programming in the first place, if a new generation was never introduced to these old ideas and therefore believed there is only one model of programming, one accepted and acceptable way of thinking about and thinking with computers. That these new generations “grow up with dogma.”
Victor mentions an incredibly important piece of education technology history in passing in his talk: PLATO (Programmed Logic for Automatic Teaching Operations), built on the ILLIAC I at the University of Illinois. PLATO, which operated out of the university’s Computer-based Education Research Laboratory (CERL) from 1960 to 1993, does represent in some ways a path that education technology (and computing technology more broadly) did not take. But if and when when its innovations were adopted (and, yes, many of them were), PLATO remained largely uncredited for its contributions.
PLATO serves in Victor’s talk as an example, along with Douglas Englebart’s NLS, of the development in the 1960s of interactive, real-time computing. In forty years time, Victor tells his imagined 1970s audience, our user interfaces will never have any delay or lag because of Moore’s Law and because “these guys have proven how important it is to have an immediately responsive UI” – a quip that anyone who’s spent time waiting for operating systems or software programs to respond can understand and chuckle remorsefully about.
This idea that computers could even attempt to offer immediate feedback – typing a letter on a keyboard and immediately seeing it rendered on a screen – was certainly new in the 1960s, as processing was slow, memory was minute, and data had to move from an input device back to a central computer and then back again to some sort of display. But the “fast round trip” between terminal and mainframe was hardly the only innovation associated with PLATO, as Brian Dear chronicles in his book The Friendly Orange Glow. That very glow was another one – the flat-panel plasma touchscreen invented by the PLATO team in 1967. There were many other advances too: the creation of time-sharing, discussion boards, instant messaging, a learning management system or sorts, and multi-user game-play, to name just a few.
The subtitle of Dear’s book – “The Untold Story of the PLATO System and the Dawn of Cyberculture” – speaks directly to his larger project: making sure the pioneering contributions of PLATO are not forgotten.
If and when PLATO is remembered (in education technology circles at least), it is as an early example of computer-assisted instruction – and often, it’s denigrated as such. Perhaps that should be no surprise – education technology is fiercely dogmatic. And it was already fiercely dogmatic by the 1960s, when PLATO was first under development. The field had, in the decades prior, developed a certain set of precepts and convictions – even if, as Victor contends in his talk at least, computing at the time had (mostly) not.
Dear begins his book where many histories of education technology do: with the story of how Harvard psychology professor B. F. Skinner had, in the late 1950s, visited his daughter’s fourth grade classroom, been struck by its inefficiencies, and argued that teaching machines would ameliorate this. The first mechanisms that Skinner built were not computerized; they were boxes with levers and knobs. But they were designed to offer students immediate feedback – positive reinforcement when students gave the correct answer, a key element to Skinner’s behaviorist theories. Skinner largely failed to commercialize his ideas, but his influence on the design of instructional machines was significant nonetheless, as behaviorism had already become a cornerstone of the nascent field of educational psychology and a widely accepted theory as to how people learn.
At its outset, the Computer-based Education Research Laboratory at the University of Illinois did not hire instructional technologists to develop PLATO. The lab was not governed by educational psychologists – behaviorists or otherwise. The programming language that was developed so that “anyone” could create a lesson module on the system — TUTOR — did not demand an allegiance to any particular learning theory. As one education professor told Brian Dear, CERL did not operate “under any kind of psychological banner. They just didn’t seem to be driven by psychological underpinnings. They were driven by a more pragmatic approach: you work with students, you work with content, you work with the technology, you put it together in a way that feels good and it will work. Whether it’s consistent with somebody’s psychology is a quickly irrelevant question.”
But it seems more likely, if we examine the history of PLATO (and perhaps even the histories of education technology and of computing technologies), that this is not really an irrelevant question at all – not in the long run at least. Certainly, the open-ended-ness of the PLATO system, as well as the PLATO culture at UI, fostered the myriad of technological innovations that Dear chronicles in The Friendly Orange Glow. But the influence of psychology on the direction of education technology – and to be clear, this was not just behaviorism, of course, but cognitive psychology – has been profound. It shaped the expectations for what instructional technology should do. It shaped the expectations for what PLATO should be. (I’d add too that psychological theories have been quite influential on the direction of computing technology itself, although I think this has been rather unexamined.)
The Friendly Orange Glow is a history of PLATO – one that has long deserved to be told and that Dear does with meticulous care and detail. (The book was some three decades in the making.) But it’s also a history of why, following Sputnik, the US government came to fund educational computing. Its also – in between the lines, if you will – a history of why the locus of computing and educational computing specifically shifted to places like MIT, Xerox PARC, Stanford. The answer is not “because the technology was better” – not entirely. The answer has to do in part with funding – what changed when these educational computing efforts were no longer backed by federal money and part of Cold War era research but by venture capital. (Spoiler alert: it changes the timeline. It changes the culture. It changes the mission. It changes the technology.) And the answer has everything to do with power and ideology – with dogma.
Bret Victor credits the message and content of his keynote to computer scientist Alan Kay, who once famously said that “the best way to predict the future is to build it.” (Kay, of course, appears several times in The Friendly Orange Glow because of his own contributions to computing, not to mention the competition between CERL and PARC where Kay worked and their very different visions of the future). But to be perfectly frank, the act of building alone is hardly sufficient. The best way to predict the future may instead be to be among those who mythologize what’s built, who tell certain stories, who craft and uphold the dogma about what is built and how it’s used.
To a certain extent, the version of “personal computing” espoused by Kay and by PARC has been triumphant. That is, PLATO’s model – networked terminals that tied back to a central machine – was not. Perhaps it’s worth considering how dogmatic computing has become about “personal” and “personalization” – what its implications might be for the shape of programming and for education technology, sure, but also what it means for the kinds of values and communities that are built without any sort of “friendly glow.”