Mar 2, 2021

PLATO (Programmed Logic for Automatic Teaching Operations) was an educational computer system that began at the University of Illinois Champaign Urbana in 1960 and ran into the 2010s in various flavors. 

Wait, that’s an oversimplification. PLATO seemed to develop on an island in the corn fields of Champaign Illinois, and sometimes precedes, sometimes symbolizes, and sometimes fast-follows what was happening in computing around the world in those decades.

To put this in perspective - PLATO began on ILLIAC in 1960 - a large classic vacuum tube mainframe. Short for the Illinois Automatic Computer, ILLIAC was built in 1952, around 7 years after ENIAC was first put into production. As with many early mainframe projects PLATO 1 began in response to a military need. We were looking for new ways to educate the masses of veterans using the GI Bill. We had to stretch the reach of college campuses beyond their existing infrastructures.

Computerized testing started with mechanical computing, got digitized with the introduction of Scantron by IBM in 1935, and a number of researchers were looking to improve the consistency of education and bring in new technology to help with quality teaching at scale. The post-World War II boom did this for industry as well. Problem is, following the launch of Sputnik by the USSR in 1957, many felt the US began lagging behind in education. So grant money to explore solutions flowed and CERL was able to capitalize on grants from the US Army, Navy, and Air Force. By 1959, physicists at Illinois began thinking of using that big ILLIAC machine they had access to. Daniel Alpert recruited Don Bitzer to run a project, after false starts with educators around the campus.

Bitzer shipped the first instance of PLATO 1 in 1960. They used a television to show images, stored images in Raytheon tubes, and a make-shift keyboard designed for PLATO so users could provide input in interactive menus and navigate. They experimented with slide projectors when they realized the tubes weren’t all that reliable and figured out how to do rudimentary time sharing, expanding to a second concurrent terminal with the release of PLATO II in 1961.

Bitzer was a classic Midwestern tinkerer. He solicited help from local clubs, faculty, high school students, and wherever he could cut a corner to build more cool stuff, he was happy to move money and resources to other important parts of the system. This was the age of hackers and they hacked away. He inspired but also allowed people to follow their own passions. Innovation must be decentralized to succeed.

They created an organization to support PLATO in 1966 - as part of the Graduate College. CERL stands for the Computer-Based Education Research Laboratory (CERL). Based on early successes, they got more and more funding at CERL. Now that we were beyond a 1:1 ratio of users to computers and officially into Time Sharing - it was time for Plato III.

There were a number of enhancements in PLATO III. For starters, the system was moved to a CDC 1604 that CEO of Control Data William Norris donated to the cause - and expanded to allow for 20 terminals. But it was complicated to create new content and the team realized that content would be what drove adoption. This was true with applications during the personal computer revolution and then apps in the era of the App Store as well. One of many lessons learned first on PLATO. 

Content was in the form of applications that they referred to as lessons. It was a teaching environment, after all. They emulated the ILLIAC for existing content but needed more. People were compiling applications in a complicated language. Professors had day jobs and needed a simpler way to build content. So Paul Tenczar on the team came up with a language specifically tailored to creating lessons. Similar in some ways to BASIC, it was called TUTOR. 

Tenczar released the manual for TUTOR in 1969 and with an easier way of getting content out, there was an explosion in new lessons, and new features and ideas would flourish. We would see simulations, games, and courseware that would lead to a revolution in ideas. In a revolutionary time.

The number of hours logged by students and course authors steadily increased. The team became ever more ambitious. And they met that ambition with lots of impressive achievements.

Now that they were comfortable with the CDC 1604 they new that the new content needed more firepower. CERL negotiated a contract with Control Data Corporation (CDC) in 1970 to provide equipment and financial support for PLATO. Here they ended up with a CDC Cyber 6400 mainframe, which became the foundation of the next iteration of PLATO, PLATO IV.

PLATO IV  was a huge leap forward on many levels. They had TUTOR but with more resources could produce even more interactive content and capabilities. The terminals were expensive and not so scalable. So in preparation for potentially thousands of terminals in PLATO IV they decided to develop their own. 

This might seem a bit space age for the early 1970s, but what they developed was a touch flat panel plasma display. It was 512x512 and rendered 60 lines per second at 1260 baud. The plasma had memory in it, which was made possible by the fact that they weren’t converting digital signals to analog, as is done on CRTs. Instead, it was a fully digital experience. The flat panel used infrared to see where a user was touching, allowing users some of their first exposure to touch screens. This was a grid of 16 by 16 rather than 512 but that was more than enough to take them over the next decade.

The system could render basic bitmaps but some lessons needed more rich, what we might call today, multimedia. The Raytheon tubes used in previous systems proved to be more of a CRT technology but also had plenty of drawbacks. So for newer machines they also included a microfiche machine that produced images onto the back of the screen. 

The terminals were a leap forward. There were other programs going on at about the same time during the innovative bursts of PLATO, like the Dartmouth Time Sharing System, or DTSS, project that gave us BASIC instead of TUTOR. Some of these systems also had rudimentary forms of forums, such as EIES and the emerging BBS Usenet culture that began in 1973. But PLATO represented a unique look into the splintered networks of the Time Sharing age.

Combined with the innovative lessons and newfound collaborative capabilities the PLATO team was about to bring about something special. Or lots of somethings that culminated in more. One of those was Notes.

Talkomatic was created by Doug Brown and David R. Woolley in 1973. Tenczar asked the 17-year old Woolley to write a tool that would allow users to report bugs with the system. There was a notes file that people could just delete. So they added the ability for a user to automatically get tagged in another file when updating and store notes. He expanded it to allow for 63 responses per note and when opened, it showed the most recent notes. People came up with other features and so a menu was driven, providing access to System Announcements, Help Notes, and General Notes. 

But the notes were just the start. In 1973, seeing the need for even more ways to communicate with other people using the system, Doug Brown wrote a prototype for Talkomatic. Talkomatic was a chat program that showed when people were typing. Woolley helped Brown and they added channels with up to five people per channel. Others could watch the chat as well. It would be expanded and officially supported as a tool called Term-Talk. That was entered by using the TERM key on a console, which allowed for a conversation between two people. You could TERM, or chat a person, and then they could respond or mark themselves as busy. 

Because the people writing this stuff were also the ones supporting users, they added another feature, the ability to monitor another user, or view their screen. And so programmers, or consultants, could respond to help requests and help get even more lessons going. And some at PLATO were using ARPANET, so it was only a matter of time before word of Ray Tomlinson’s work on electronic mail leaked over, leading to the 1974 addition of personal notes, a way to send private mail engineered by Kim Mast.

As PLATO grew, the amount of content exploded. They added categories to Notes in 1975 which led to Group Notes in 1976, and comments and linked notes and the ability to control access.

But one of the most important innovations PLATO will be remembered for is games. Anyone that has played an educational game will note that school lessons and games aren’t always all that different. Since Rick Blomme had ported Spacewar! to PLATO in 1969 and added a two-player option, multi-player games had been on the rise. They made leader boards for games like Dogfight so players could get early forms of game rankings. Games like airtight and airace and Galactic Attack would follow those.

MUDs were another form of games that came to PLATO. Collosal Cave Adventure had come in 1975 for the PDP, so again these things were happening in a vacuum but where there were influences and where innovations were deterministic and found in isolation is hard to say. But the crawlers exploded on PLATO. We got Moria, Oubliette by Jim Schwaiger, Pedit5, crypt, dungeon, avatar, and drygulch. We saw the rise of intense storytelling, different game mechanics that were mostly inspired by Dungeons and Dragons, As PLATO terminals found their way in high schools and other universities, the amount of games and amount of time spent on those games exploded, with estimates of 20% of time on PLATO being spent playing games. 

PLATO IV would grow to support thousands of terminals around the world in the 1970s. It was a utility. Schools (and even some parents) leased lines back to Champagne Urbana and many in computing thought that these timesharing systems would become the basis for a utility model in computing, similar to the cloud model we have today. But we had to go into the era of the microcomputer to boomerang back to timesharing first. 

That microcomputer revolution would catch many, who didn’t see the correlation between Moore’s Law and the growing number of factories and standardization that would lead to microcomputers, off guard. Control Data had bet big on the mainframe market - and PLATO. CDC would sell mainframes to other schools to host their own PLATO instance. This is where it went from a timesharing system to a network of computers that did timesharing. Like a star topology. 

Control Data looked to PLATO as one form of what the future of the company would be. Here, he saw this mainframe with thousands of connections as a way to lease time on the computers. CDC took PLATO to market as CDC Plato. Here, schools and companies alike could benefit from distance education. And for awhile it seemed to be working. Financial companies and airlines bought systems and the commercialization was on the rise, with over a hundred PLATO systems in use as we made our way to the middle of the 1980s. Even government agencies like the Depart of Defense used them for training. But this just happened to coincide with the advent of the microcomputer.

CDC made their own terminals that were often built with the same components that would be found in microcomputers but failed to capitalize on that market. Corporations didn’t embrace the collaboration features and often had these turned off. Social computing would move to bulletin boards And CDC would release versions of PLATO as micro-PLATO for the TRS-80, Texas Instruments TI-99, and even Atari computers. But the bureaucracy at CDC had slowed things down to the point that they couldn’t capitalize on the rapidly evolving PC industry. And prices were too high in a time when home computers were just moving from a hobbyist market to the mainstream. 

The University of Illinois spun PLATO out into its own organization called University Communications, Inc (or UCI for short) and closed CERL in 1994. That was the same year Marc Andreessen co-founded Mosaic Communications Corporation, makers of Netscape -successor to NCSA Mosaic. Because NCSA, or The National Center for Supercomputing Applications, had also benefited from National Science Foundation grants when it was started in 1982. And all those students who flocked to the University of Illinois because of programs like PLATO had brought with them more expertise.

UCI continued PLATO as NovaNet, which was acquired by National Computer Systems and then Pearson corporation, finally getting shut down in 2015 - 55 years after those original days on ILLIAC. It evolved from the vacuum tube-driven mainframe in a research institute with one terminal to two terminals, to a transistorized mainframe with hundreds and then over a thousand terminals connected from research and educational institutions around the world. It represented new ideas in programming and programming languages and inspired generations of innovations. 

That aftermath includes:

• The ideas. PLATO developers met with people from Xerox PARC starting in the 70s and inspired some of the work done at Xerox. Yes, they seemed isolated at times but they were far from it. They also cross-pollinated ideas to Control Data. One way they did this was by trading some commercialization rights for more mainframe hardware. 
• One of the easiest connections to draw from PLATO to the modern era is how the notes files evolved. Ray Ozzie graduated from Illinois in 1979 and went to work for Data General and then Software Arts, makers of VisiCalc. The corporate world had nothing like the culture that had evolved out of the notes files in PLATO Notes. Today we take collaboration tools for granted but when Ozzie was recruited by Lotus, the makers of 1-2-3, he joined only if they agreed to him funding a project to take that collaborative spirit that still seemed stuck in the splintered PLATO network. The Internet and networked computing in companies was growing, and he knew he could improve on the notes files in a way that companies could take use of it. He started Iris Associates in 1984 and shipped a tool in 1989. That would evolve into what is would be called Lotus Notes when the company was acquired by Lotus in 1994 and then when Lotus was acquired by IBM, would evolve into Domino - surviving to today as HCL Domino. Ozzie would go on to become a CTO and then the Chief Software Architect at Microsoft, helping spearhead the Microsoft Azure project.
• Collaboration. Those notes files were also some of the earliest newsgroups. But they went further. Talkomatic introduced real time text chats. The very concept of a digital community and its norms and boundaries were being tested and challenges we still face like discrimination even manifesting themselves then. But it was inspiring and between stints at Microsoft, Ray Ozzie founded Talko in 2012 based on what he learned in the 70s, working with Talkomatic. That company was acquired by Microsoft and some of the features ported into Skype. 
• Another way Microsoft benefited from the work done on PLATO was with Microsoft Flight Simulator. That was originally written by Bruce Artwick after leaving the university based on the flight games he’d played on PLATO. 
• Mordor: The Depths of Dejenol was cloned from Avatar
• Silas Warner was connected to PLATO from terminals at the University of Indiana. During and after school, he wrote software for companies but wrote Robot War for PLATO and then co-founded Muse Software where he wrote Escape!, a precursor for lots of other maze runners, and then Castle Wolfenstein. The name would get bought for $5,000 after his company went bankrupt and one of the early block-buster first-person shooters when released as Wolfenstein 3D. Then John Carmack and John Romero created Doom. But Warner would go on to work with some of the best in gaming, including Sid Meier.  
• Paul Alfille built the game Freecell for PLATO and Control Data released it for all PLATO systems. Jim Horne played it from the PLATO terminals at the University of Alberta and eventually released it for DOS in 1988. Horn went to work for Microsoft who included it in the Microsoft Entertainment Pack, making it one of the most popular software titles played on early versions of Windows. He got 10 shares of Microsoft stock in return and it’s still part of Windows 10 using the Microsoft Solitaire Collection..
• Robert wood head and Andrew Greenberg got onto PLATO from their terminals at Cornell University where they were able to play games like Oubliette and Emprie. They would write a game called Wizardry that took some of the best that the dungeon crawl multi-players had to offer and bring them into a single player computer then console game. I spent countless hours playing Wizardry on the Nintendo NES and have played many of the spin-offs, which came as late as 2014. Not only did the game inspire generations of developers to write dungeon games, but some of the mechanics inspired features in the Ultima series, Dragon Quest, Might and Magic, The Bard’s Tale, Dragon Warrior and countless Manga. Greenberg would go on to help with Q-Bert and other games before going on to work with the IEEE. Woodhead would go on to work on other games like Star Maze. I met Woodhead shortly after he wrote Virex, an early anti-virus program for the Mac that would later become McAfee VirusScan for the Mac.
• Paul Tenczar was in charge of the software developers for PLATO. After that he founded Computer Teaching Corporation and introduced EnCORE, which was changed to Tencore. They grew to 56 employees by 1990 and ran until 2000. He returned to the University of Illinois to put RFID tags on bees, contributing to computing for nearly 5 decades and counting. 
• Michael Allen used PLATO at Ohio State University before looking to create a new language. He was hired at CDC where he became a director in charge of Research and Development for education systems There, he developed the ideas for a new computer language authoring system, which became Authorware, one of the most popular authoring packages for the Mac. That would merge with Macro-Mind to become Macromedia, where bits and pieces got put into Dreamweaver and Shockwave as they released those. After Adobe acquired Macromedia, he would write a number of books and create even more e-learning software authoring tools. 

So PLATO gave us multi-player games, new programming languages, instant messaging, online and multiple choice testing, collaboration forums, message boards, multiple person chat rooms, early rudimentary remote screen sharing, their own brand of plasma display and all the research behind printing circuits on glass for that, and early research into touch sensitive displays. And as we’ve shown in just a few of the many people that contributed to computing after, they helped inspire an early generation of programmers and innovators. 

If you like this episode I strongly suggest checking out The Friendly Orange Glow from Brian Dear. It’s a lovely work with just the right mix of dry history and flourishes of prose. A short history like this can’t hold a candle to a detailed anthology like Dear’s book. 

Another well researched telling of the story can be found in a couple of chapters of A People’s History Of Computing In The United States, from Joy Rankin. She does a great job drawing a parallel (and sometimes direct line from) the Dartmouth Time Sharing System and others as early networks. And yes, terminals dialing into a mainframe and using resources over telephone and leased lines was certainly a form of bridging infrastructures and seemed like a network at the time. But no mainframe could have scaled to the ability to become a utility in the sense that all of humanity could access what was hosted on it. 

Instead, the ARPANET was put online and growing from 1969 to 1990 and working out the hard scientific and engineering principals behind networking protocols gave us TCP/IP. In her book, Rankin makes great points about the BASIC and TUTOR applications helping shape more of our modern world in how they inspired the future of how we used personal devices once connected to a network. The scientists behind ARPANET, then NSFnet and the Internet, did the work to connect us. You see, those dial-up connections were expensive over long distances. By 1974 there were 47 computers connected to the ARPANET and by 1983 we had TCP/IPv4.And much like Bitzer allowing games, they didn’t seem to care too much how people would use the technology but wanted to build the foundation - a playground for whatever people wanted to build on top of it.

So the administrative and programming team at CERL deserve a lot of credit. The people who wrote the system, the generations who built features and code only to see it become obsolete came and went - but the compounding impact of their contributions can be felt across the technology landscape today. Some of that is people rediscovering work done at CERL, some is directly inspired, and some has been lost only to probably be rediscovered in the future.  One thing is for certain, their contributions to e-learning are unparalleled with any other system out there. And their technical contributions, both in the form of those patented and those that were either unpatentable or where they didn’t think of patenting, are immense. 

Bitzer and the first high schoolers and then graduate students across the world helped to shape the digital world we live in today. More from an almost sociological aspect than technical. And the deep thought applied to the system lives on today in so many aspects of our modern world. Sometimes that’s a straight line and others it’s dotted or curved. Looking around, most universities have licensing offices now, to capitalize on the research done. Check out a university near you and see what they have available for license. You might be surprised. As I’m sure many in Champagne were after all those years. Just because CDC couldn’t capitalize on some great research doesn’t mean we can’t.