The Friendly Orange Glow

by Brian Dear

  The new company would be called Iris Associates. “We were completely independent,” says Ozzie, “we owned all rights to what we did, except they had, in exchange for providing all this funding, an option that they could exercise that would essentially, if they exercised the option, they committed to paying us a certain royalty rate. They got exclusive marketing rights—actually, they basically got all rights to the product and we got compensated by royalties—but they would have to use best efforts to market the product. They couldn’t buy it and sit on it. Year after year went by and they just kept pumping more and more money but they knew that I knew Gates very well. So they knew that if they stopped funding it, I would immediately take the product, ’cause I owned it, to Microsoft. So it was sort of like they’re damned if they do, damned if they don’t, so for many years while they couldn’t figure out how to market the thing, they just kept funding it, figuring it’s better to keep it going than to give it up. Eventually in January of ’88 they exercised the option and it became theirs, essentially the product became theirs at that point, and they were committed to marketing it and they were committing to an affirmative obligation to pay royalties, and so we still operated independently collecting royalties through ’94.”

  Jim Manzi characterizes the Lotus-Iris deal as “very unusual, and it largely worked not because of the brilliance in the contracting language, but I think it largely worked because of the relationship between the principals over time.”

  Lotus Notes, the official name of the product when it finally shipped, offered email, calendaring and scheduling, an address book, access lists, document commenting, online forums, anonymous notes, the equivalent of a Notesfile Sequencer, a database, and programming tools to build custom applications within the Notes environment. The Iris team took a pile of PLATO ideas they’d lived and breathed at CERL and transferred them into a Microsoft Windows environment for the PC. But however impressive the final product, it was the kind of tool that required an entire organization to be trained on and commit to—it didn’t work if only small clusters of employees used it. That meant an entire organization had to change their behavior and reengineer itself in order to fully exploit Notes’s features. Lotus decided that even though the product was for workgroups, it was not going to work well for small workgroups—who would install it? Who would administer it? No, it was better suited for an enterprise. To make that abundantly clear to the marketplace, the company set the starting price for the product at $64,000. Their first customer was Price Waterhouse, who were so impressed with the product they ordered a historic ten-thousand-user license, the largest single order for a software program in the computer industry up to that time. Other corporations soon followed with their own orders.

  Notes was a hit.

  Ozzie attributes at least part of Notes’s success in the enterprise space to serendipity: “In the late 1980s this country’s economy was going through an odd transition, people were being laid off and there was this intense amount of restructuring, under the auspices of reengineering, that was starting to happen. And we were available right at that time as a tool that people felt could help reengineering.”

  Lotus finally acquired Iris Associates in 1994 for approximately $84 million, ten years after Ozzie, Kawell, and Halvorsen had launched the tiny firm with a dream of bringing a major chunk of PLATO to the wider world. Across that decade Lotus had undergone a stunning transformation from the company known for 1-2-3 to becoming the company known for Notes. The timing was fortuitous, because Microsoft had aggressively been out to kill 1-2-3 with its own product, Excel, which had by then become and has remained to today the leading spreadsheet in the market. Lotus Notes meanwhile became so attractive to large enterprises, and, once installed enterprise-wide, so difficult to migrate away from (a desirable situation venture capitalists term a “high switching cost”), that IBM decided to buy Lotus in its entirety in July 1995 for $3.52 billion.

  The collaborative, group-productivity ideas embodied in PLATO—tools largely created by PLATO’s own users, many of whom were teenagers at the time—finally had an impact on the world. It would be the largest financial transaction in history for a project deeply inspired by PLATO and developed by PLATO veterans, and, ironically, it had nothing to do with computer-based education. In time IBM’s Notes would surpass 120 million users. Few of them would ever know of the product’s deep connection to PLATO.

  Ozzie formed a new company in 1997, Groove Networks, which took the ideas of Lotus Notes and reapplied them using a more decentralized, peer-to-peer fashion. Groove was acquired by Microsoft in 2005 for $120 million. Ozzie became the chief software architect of Microsoft, laying the groundwork for the Azure project, which became Microsoft Cloud. In 2010 he left and formed Talko (named after Talkomatic), a mobile conferencing platform largely focused on voice messages and annotations to those messages. In 2015 Microsoft acquired Talko, which the company absorbed into Skype, the global Internet communications service Microsoft had acquired in 2011.

  Epilogue

  “Service interruption in 60 seconds…”

  The familiar message appeared at the bottom of the computer screens of the twenty-two die-hard NovaNET users scattered around the country. It was 2:16 a.m. Central Time on the first of September, 2015. For these night owls, these service interruptions were a part of life, as predictable as the sun rising and setting. If you had been a user of the system for decades, possibly since you were a kid, a service interruption provided a welcome break, and sometimes the prospect of an exciting change, perhaps a new system feature or TUTOR command, when the system was brought back up.

  What was different, this time, was that the service was not being interrupted. It would not be brought back up. It was being terminated. Here they were, holding out in the original cyberspace, the cyberspace that had originally been powered by computers actually called CYBERs, gathered to witness the termination of the last, longest-lasting, virtual place they’d called home for decades. This time the system was going to kick them off for good and then die.

  Some of these users had stayed up late at night ostensibly to make one final set of backups of their old work dating back possibly to the 1970s. Others were there to hang out, see who showed up for the end of the world as they’d known it, and post a few messages to the others, knowing, with the end nigh, that no one would see these postings ever again. One user described the scene as “a band of fogeys who collected to roast marshmallows over the flames.”

  Others described how long ago they had gotten their start on PLATO, and continued using NovaNET over the years. “I just calculated,” said one user, “the number of days from my first encounter with PLATO III to today: 16,285 days.”

  “Never found a better online community,” said another. “One or two, almost as good. All things end, and unfortunately it is time for NovaNET to end.”

  This wasn’t going to be like a TV station going off the air, the familiar script that begins with scratchy old footage of an American flag waving in the wind against a blue sky, perhaps a fighter jet flying by, while an orchestra played “The Star-Spangled Banner” a little faster than usual, followed by the issuance of a final station identification message, then a video color bar test pattern and a shrill audio tone thrown in for good measure, and then, click, off the air, followed by the video snow and audio static signifying nothingness, as the viewer gently dozed off. No, this was NovaNET, the computer-based education service and online community, and the twenty-two diehards staying up this late were invariably men, mostly white-haired, graybeard ghosts who quietly haunted one of the oldest and least known continuously operating computer networks in the world, going on some thirty-five years or forty-five years, depending on who was debating the NovaNET vs. PLATO IV origins. Some of these old diehards seemed to hang around online for no discernible reason, drawn perhaps out of nostalgia or loneliness, like lone regulars at a local neighborhood saloon with its flickering, fading neon beer sign out front and a handful of customers inside
quietly drinking away the hours, days, and years.

  NovaNET’s official mission had always been educational, delivering self-paced, individualized instruction to students primarily situated in troubled schools or remedial education programs around the country: students who had not been successful in normal schools, but found a way to finish their high school diploma, or achieve a GED certificate, in an institution that offered the private one-on-one ever-patient digital tutoring in front of a NovaNET terminal or a Mac or a PC running the NovaNET software. An intimate, quiet setting—not dissimilar to CERL’s and CDC’s PLATO learning centers in prisons that had been successful for so many years—a setting where the student could focus on his studies and not worry about the pressure of peers or the disappointed gaze of a harried instructor. For years NovaNET had successfully delivered on its mission, educating hundreds of thousands—more likely millions all totaled—of students all over the country, helping them get on with their lives, helping them get better jobs, better incomes, enough to support a family.

  But NovaNET had another community of users, and it was this tiny group that was there to bear witness to the end, the closing of a system whose origins go back fifty-five years. A system whose vision, both from educational and technical implementation standpoints, had not changed since day one. This other small group of users were the authors, the systems programmers, and former gamers and participants in the online social community that boomed in the 1970s and stayed active until the early 1990s, after which only the diehards would stop by each evening to check in, say hello, offer a quip or two, and sign off. Tonight was the ultimate sign-off: they were the final few, perched on the topmost point of the last Ziggurat, experiencing the final seconds of NovaNET’s existence.

  —

  So much had happened since CDC had sold off PLATO in 1989, and CERL had shut down a few years later. The Roach Organization, which had bought the rights to PLATO from CDC, eventually renamed itself TRO Learning, went public (its stock symbol, TUTR, a knowing wink to PLATO’s programming language), then renamed itself again to PLATO Learning. In 2012, in what seemed an admission that the PLATO name had become an embarrassing reference to a past the marketplace now knew nothing about, the company renamed itself once again, this time as Edmentum.

  VCampus, one of the spinoffs of CDC’s PLATO business, was long gone by 2015. Drake, the testing company, was acquired in 1999 by National Computer Systems, which then soon acquired NovaNET Learning. Pearson, the giant British education conglomerate, then bought NCS in 2000 for $2.5 billion. NovaNET, reputed to be the only profitable division within Pearson in the early 2000s, quietly blossomed, reaching a peak number of simultaneous users, 13,184, on June 18, 2008. That would turn out to be NovaNET’s high-water mark. By 2010 the whispers were growing louder that NovaNET’s days were numbered, and in October 2014, the news became official: Pearson told customers that NovaNET would shut down in 2015, and everyone would need to wrap things up over the next academic year.

  Since the close of CERL, the world had been transformed by the Internet and the World Wide Web. During the PLATO era most people around the world had no idea what computer-based education was, or what an online community was, let alone the benefits of email, chat rooms, instant messaging, notesfiles, and multiplayer games. In an astonishingly short period of time in the mid-1990s, hundreds of millions of people would suddenly not only discover these technologies, but would soon not be able to imagine how they could live their daily lives without them.

  Computer-based education—a dinosaur-era term not used anymore, now known by various buzzwords, including e-learning—is booming more than ever. Millions of people study courses online around the world. Massive open online courses (MOOCs) exploded on the scene after the Kahn Academy pioneered the use of educational videos on YouTube. In 2013, the CEOs of the two leading MOOC services, Udacity and Coursera, were asked if they had ever heard of the PLATO system. “No,” both said, blank expressions on their faces.

  Schools continue to spend billions on computers, software, and networks, but the question remains who is benefiting more, students or the vendors. Educational technology has become a hot area for venture capitalists, who now routinely pour billions of dollars per year into new start-ups. Much of the technology developed by CERL in the 1970s has never been replicated.

  Part of the CERL dream was to build a system so flexible, so infinitely capable, with such a vast array of tools, that any educator, regardless of skill level, could sit down and author his or her own material for students. This model worked at the University of Illinois and the University of Delaware, and found some traction at other universities and large institutions during CDC’s peak years with PLATO. Michael Allen took the idea to the Macintosh with Authorware, launching the next generation of powerful authoring tools designed to give any educator or trainer the ability to craft whatever they had in mind to teach their students. But even Authorware is no more; Adobe discontinued it years ago. There are still dozens of authoring systems today, but the notion of spending time and money to develop custom courseware using programming tools has lost almost all of its early energy, and only big corporations and governments have budgets that can justify custom development. The costs of developing one hour of courseware—actual interactive lessons developed on the computer and then used by the student via the computer—never came down. If anything, they have continually climbed since the 1970s. MOOCs are, in a way, a reaction to this authoring quandary: educators seem to have punted on the whole question of devoting resources to creating interactive materials the way professors like Stan Smith so eagerly—and brilliantly—did on PLATO. Instead, they plop a video camera down in front of the lecture hall, and record a lecture to students, then toss it on YouTube or one of the MOOC websites. This became the new definition of “courseware” after 2010. MOOC companies and educators are working to provide more sophisticated forms of interaction beyond students passively watching videos and then going through pages of multiple-choice questions, but they’re starting from scratch: today most educators have never heard of PLATO or TUTOR, or that anything that powerful could have existed before many of them were born. PLATO has even faded from the institutional memory of the University of Illinois; professors there, like everywhere else, are trying all sorts of new tools to deliver education online, running into problems that were once solved decades ago. The field of educational technology, largely ignorant of its own history, seems eternally condemned to repeat itself.

  If CDC had tried harder, could they have gotten millions of terminals in the schools? Bob Morris is doubtful. “While we were very ambitious about getting the cost of delivering an hour’s worth of PLATO into the schools, it was still expensive. My belief is that we had gotten it down to the point where it was reasonable. However, what’s reasonable to us is not necessarily to the customer. One of the things that we always fought with PLATO was we would bring people in from academia, whether it would be universities, primary schools, and that sort of thing, and give them a full briefing on PLATO, and by the time they left, they would be so excited they couldn’t see straight. And you’re convinced that within a week you’re going to get an order. Well, the problem kept being that they had alternatives for using the money. And the school administrators are rewarded more for the size of their staff than they are for the technology they bring in. And if PLATO services or a PLATO system would cost them so much, they would say, ‘How much bigger staff can I get by spending that same money?’ And invariably we lost to their increasing their staff or their not reducing their staff.”

  Walter Bruning, a CDC executive, once explained to reporters in the mid-1980s what William Norris’s deep, underlying motivation with PLATO was really about. “His fundamental philosophy is to operate the schools,” he said. “He cares a lot about PLATO, the name. He cares a lot about courseware and what we have learned (about education technology). But when Bill talks about educational delivery, what he is really talking about is running the schools….We are
on a novel and maybe even revolutionary path. We are after the privatization of one of the largest public services in this country—the privatization of the public schools.”

  Since that time, the notion no longer seems so outlandish. As Mark Andreessen has famously said, “Software is eating the world.” It seems that big business, including Silicon Valley, is more determined than ever to devour public education and turn it into a monetized business. Whether the passionate, caring, hands-on Mary Graves of the world have a place in a world of private, for-profit schools, no doubt loaded to the hilt with technology, remains to be seen.

  —

  “Tick tick tick,” one of the last-stand NovaNET users wrote at 1:52 a.m. NovaNET normally shut down at 2:00 a.m. each night for maintenance, and there was no reason to doubt tonight was not like any other night. But at 2:00 a.m., the system was still running. August passed into September, and NovaNET was still up, twenty-two stubborn users still online and waiting for the end. “As far as the operator is concerned,” John Hegarty, one of the last remaining “s” programmers on the Pearson payroll, posted at 2:01 a.m., “it is the same as ever. They don’t know what’s going on.” Somewhere in Iowa, at Pearson’s data center, an unwitting system operator seemed to have forgotten to press a few buttons to shut the system down.

  A small contingent of the new wave, now decidedly old wave, waited for the end of an era, scheduled for midnight on August 31, 2015, and the end was now several hours late.

  “I hope I don’t have to use the console,” John Hegarty said at 2:04 a.m. “I always hated that thing.”