The Friendly Orange Glow

by Brian Dear

  Tenczar was no Blomme. His reaction: furious. “He hated that whole thing,” says Bloomfield, “and he basically just beat on me to just tell him how I cracked security and I’ll fix it. Okay, so I told him, it was using a -jumpout- command, I jumped right into the middle of, I don’t know, was it the ‘edit’ program or something like that, and I managed to set my password to anything I wanted to, including things you could not type in.”

  Bloomfield had discovered that system passwords contained untypeable characters. “You can’t type in 0 characters, right? They didn’t exist,” he says. “So I could jump in there, and the way in which they handled system privileges was, they put passwords on the lessons that they did not want any normal person to be able to enter, they would put a password that you couldn’t type, it would just like say, ‘system,’ but it would be zero zero zero, the zero character and then ‘system’ at the right, something like that, and I would just jump there with that, with my variable set to that, and it was as though I typed in an untypeable value, it would accept my password, and I was in business.”

  Tenczar remained furious, but Blomme chuckled at Bloomfield’s exploit. “He thought this was wonderful,” says Bloomfield. Blomme realized Bloomfield had discovered a fundamental flaw in the way CERL’s systems staff had set up password security on PLATO. The passwords were not secure enough, and there was a way to circumvent them due to an architectural quirk in CDC computers, a quirk that the teenage Bloomfield had figured out on his own. For Blomme, this was just the kind of thing he admired in this new wave of bright kids hacking away at PLATO.

  —

  One reason these Big Board games were so popular, in addition to the social aspect of playing against someone else, was that they were so playable. “Playability” also meant they were simple and lightweight, with rudimentary Xs and Os or tiny little spaceships or other tokens. The real guarantee of playability goes all the way back to the diagram of PLATO I: Bitzer’s Fast Round Trip architecture, which required that a PLATO user could type something at her keyboard and the letter she typed would show up on her screen within about one-tenth of a second. PLATO, the ever-patient automatic teacher, was also supposed to be ever responsive. All through these years, from 1960 right up until the rollout of PLATO IV in 1972 and onward, the Fast Round Trip rule defined the PLATO experience. With the new multimillion-dollar PLATO IV system, things were no different. When the gamer kids discovered the Fast Round Trip in the late 1960s, and then armies of them rediscovered it in the early 1970s, it provided an unbeatable platform for fast—indeed, frantic—games that required not only skill but sheer keyboard speed to survive and thrive. All of the Big Board games, as did practically every other PLATO-designed game, took advantage of the Fast Round Trip. It was a key part of what made a PLATO game a PLATO game.

  —

  Bloomfield briefly became PLATO’s game author celebrity, a hit-maker extraordinaire that everyone knew as “louisb.” First Moonwar, and then Mazewar, another Big Board favorite, this time where the trusty X and O started out on different sides of an on-screen maze, and you had to try to get to the other side before your opponent got to your side.

  Bloomfield wasn’t the only one interested in writing games. Once it became known that one could write a game, everybody wanted to write a game. But not everybody had connections yet, a sponsor or mentor the likes of Blomme or even Tenczar. Most of these kids were gamers, thrill seekers, staying up all night in the crazy halls of CERL during 1972 through 1974. “CERL was a zoo in those days,” recalls Erik Witz, another gamer kid addicted to the PLATO scene. “Gaming was rampant. Pizza boxes and other trash was stacked to the ceilings. The most popular games where all Louis Bloomfield’s, I believe….Then CERL cracked down. Everyone was required to have user IDs. Before that you could log on just by pressing NEXT to begin. The next day I hacked into the system simply by trying common names and hitting one that brought up an initial password page. A week later I’d given IDs to all my friends. The ops were astounded when they walked in and saw all the kids still playing games after they had implemented the security.”

  In a way, there were multiple levels of games going on. At the base level were the actual PLATO games that you could see happening and becoming ever more widespread. But at a more meta-level was PLATO itself, or at least the act of using it—after first figuring out how to get access to it—and in a way that was becoming a game. Initially, gaming at CERL was unrestricted. The numbers were manageable. The drain on the system was tolerable. But as resources became tighter, the number of complaints began to rise, the number of parents calling or coming by wondering where the hell their teenage son was kept growing, and in general things began to get more and more out of hand. The PLATO Services Organization online consultant staff on the second floor, and the system operators on the first floor, had both unwittingly become Os in their own right, fighting the gamers’ Xs. The operators were determined to gain the upper hand, upping the ante in this growing competition with ever-more-ingenious gamer kids. This forced further development of restrictions to access in various PLATO classrooms, banning unauthorized use outright, or more likely limiting use to certain hours of the day. Over the next few years, the CERL staff would need to resort to even more strict rules, including creating lists of games, notesfiles, and other recreational files that were again banned outright or during certain hours. They also built tools so that any site on the network (other universities, government installations, etc.) could manage their own rules for controlling access to the terminals and phone lines for which they paid considerable monthly fees.

  “That summer it kept getting more and more brutal to game unrestrictedly,” says one former PLATO gamer we’ll call Sam. “Mostly because of the efforts of people like myself, who were there all day and half the night. I had a clearance from the U of I cops right next door, in those days, to be out after curfew. I’d ride my little bike over to CERL, lock the bike up, and go in and game, you know, maybe break for dinner, then stay until two to three or overnight sometimes.”

  —

  Doug Green remembers that when he got started, most of the login security had yet to be implemented. “You just walk into PLATO,” he says, “press NEXT to begin, and they asked for name and course [later changed to ‘group’], but it meant nothing. You could type anything in. It was just a pass-through. They were in the process of attempting to figure out how to implement security and the argument was, Well, do we really need this, you know, it’s all open right now, it’s all community access right now, we want to keep it that way, but…too many people are destroying the commons, and so we got to control this thing.”

  One night in the early spring of 1973, Green arrived at CERL, “looking forward to a night of domination on the Big Boards of Moonwar and Mazewar, for I was becoming quite proficient at each,” he says. “But stunningly, the room was empty. Some people I knew were sitting forlornly at various terminals, unable to proceed. A new security system had been implemented just that day. From now on, to use PLATO you had to have a ‘name’ and that name had to be registered in a ‘course’ and you had to have a ‘password’ that matched the name and the course. A few people were actually ‘logged on’ and were—gasp—working….But the game players were shut out. And so was I. Crushed. Angry. Frustrated. And determined to find a way around the predicament.”

  The few, the sad, the gamers. Those unlucky few who stuck around the library-quiet PLATO terminal room simply could not tear themselves away, even if they couldn’t get online. But was it really impossible to get on? The “there-were-many-ways-to-get-around-this” wisdom soon spread among the most determined gamers, including Green. “Sitting there pounding away at the keyset, anxiously and futilely, one of us suddenly whispered, ‘Hey, I’m in.’ And there he was, at the User Mode screen, somehow suddenly able to access games and all, just as before. How had he done it? Even he didn’t know. We experimented, and experimented some more. And then we learned.” It turned out that if you pressed NEXT
to begin, like the screen told you, then pressed BACK on the next screen that asked for your user name, returned to the “Press NEXT to begin” screen and pressed NEXT, then BACK, then NEXT, then BACK, in rapid succession, something broke. “You could fill the keyboard buffer and overload your timeslice allocation,” Green says. “And the timeslice overload error would bomb you right past the security system and onto the User Mode page.”

  The hours ticked by, and people straggled away one by one. Only Green was left, mesmerized by things he was finding by simply poking around the wide-open system. “I had the entire room to myself. And I had found a brand-new course whose owner had not yet set up his passwords….And so I performed my first hack on a computer system….I knew I was doing wrong, but I was desperate. So I put my own password on that course, and I set up enough names and passwords to log on every terminal in the room. I had every terminal doing something, anything, drunk with power and showing off to myself how cool I was. And as I sat at one terminal, making entries in my notebook where I catalogued the lessons on PLATO, I looked into the screen and saw the reflection of Rick Blomme standing in the doorway behind me, scratching his head. He must have been working late and checked to see who was on the system, and seen the anomaly. I was scared, but played it cool like he wasn’t there, just kept on making notes in my notebook. Rick scratched his head a moment, then left and returned a few minutes later with another systems programmer whose name I didn’t know. They stood in the doorway together, scratching their heads and muttering to themselves. I watched them in the reflection on my screen, pretending to be making notes. The other programmer left and Rick stood there alone, fingering his beard. Finally I couldn’t take the suspense anymore. I turned my chair around and looked up and said, ‘Hi, Rick,’ and smiled a Cheshire Cat grin at him. He slowly walked over and sat down across from me and just sort of stared at me for a while.”

  The conversation, as Green recalls it, proceeded thus:

  BLOMME: Uh, what, er, um, how…uh…did you do this?

  GREEN: But Rick, if I tell you, you’ll fix it and I won’t be able to get on anymore.

  (At this point Green recalls Blomme giving him “one of those long penetrating stares” before speaking.)

  BLOMME: Okay, okay…suppose I give you records in one of my courses?

  GREEN: Well, okay, uh, what about some file space? I’d like to try some programming.

  “And he stared at me and slowly began to nod his head,” Green says. “And that’s how I got onto PLATO, legitimately….Within days I was learning to program. I’d earned my pilot’s license the year before, and my first program was a lesson on the fundamental forces of flight.” Blomme was impressed, and when Green graduated high school in the spring of 1973, Blomme recommended him to the Aviation Research Lab (ARL). He wound up getting a job there that summer. But he kept coming back to CERL. And he started working on a game of his own, influenced by all the aviation stuff he was doing over at ARL: Dogfight.

  In the new Big Board game Dogfight, instead of an X and an O, there were two tiny airplanes on the screen, going at it, trying to shoot each other down. PLATO’s limitations meant it was impossible to render 3D graphics and fancy photorealistic animations like ones we see today, so Green kept it simple, the way Blomme and Bloomfield had done in their respective games. He made it a top-down two-dimensional game, and limited the graphics to two tiny representations of airplanes, just like Spacewar had two tiny spaceships.

  Dogfight went viral; people loved it. “For a while it was the most popular game on the system, racking up more hours of play than even Moonwar. I wrote the original version using trigonometry to figure the angles and directions, then Larry White modified it to use a much faster table look-up scheme.”

  Larry White was another one of these brilliant Illinois students who wound up at CERL as a systems programmer. Green held him in awe. “Anytime I got stuck with a coding problem or especially debugging,” Green says, “he was so good at debugging, you know, he could just glance, oh here, a comma out of place.”

  The PLATO system during this era was constantly crashing. Unlike today, where the software companies routinely write new code on a test machine, not the live production machine, during the mainframe era of PLATO, code was written on the live production machine—the only machine CERL had—and then tested at night while the expendable gamers were in force. Gamers were tolerated at night because the games they played often utilized more TUTOR functions than conventional educational lessons, and with the games being pounded upon all night long, if a systems programmer made a mistake while improving or changing a TUTOR command, it could very well show up as a crashed program or even a crashed system in very short order.

  One time there was a critical system bug, crashing PLATO “deader than a doornail,” Green recalls. “And they could not find it. Rick Blomme and Larry White wrote a specific program to look for this bug. How they did it, I don’t know—but when this bug occurred, they dumped the entire core [the CYBER’s main memory] out to the big line printer.” Green walked into the machine room on the fourth floor, saw that the PLATO system had crashed, and saw Blomme and White with the pile of “core dump” printouts. “Larry and Rick are sitting side by side and they are scanning down, turn the page, and one set of eyeballs is scanning left half, and the other set of eyeballs is scanning right half….Insane, trying to look something up. So imagine, just a string of zeros and ones and ABCDEFG’s and that’s all it is and as fast as you could turn the page, they’d turn the page, and scan down, turn the page, and scan down, and turn the page, and scan down, and turn the page, and scan down. It took them all night long to find out one bit out of position. And the next day it was fixed.”

  In Dogfight White used the pseudonym “Jackal” (after the film The Day of the Jackal). Blomme went by “Red Baron.” Another PLATO gamer named David Frye called himself “Fright Pilot.” Even Don Bitzer played the game, and always went by the name “pig,” according to White. A feature of Dogfight’s Big Board was that it checked the signon of each player as they entered the program. If the signon matched an entry in a special list, then that user’s preset pseudonym was automatically entered, saving them a step. “This allowed faster entry for the authors of the game and some of their friends,” says White. One day one of Dogfight’s authors added Don Bitzer to the special list so that his pseudonym “pig” was always available to him. “A few weeks later,” says White, “I learned from Don that this was a problem for him. Don would give demos to lots and lots of people, and one of the demos he often did would be to demonstrate the graphics and interactive and inter-terminal capabilities of PLATO using a game program. Part of his presentation was about how each player could select their own game name—but suddenly he could no longer do this in Dogfight and thus he could no longer demonstrate one of the social aspects of inter-personal interaction on PLATO.” Bitzer’s entry in the list was quickly removed.

  The Dogfight authors thought up an additional trick that other game authors soon copied. If you used the TUTOR -slide- command, you could instruct your terminal to select one of the 256 microfiche slide images that would then be projected through the glass of the plasma display and out to the user, with any orange text and graphics magically superimposed. At least, that was the intent. But leave it to the game authors to find another use: special effects. No need to insert a microfiche sheet into the machine. The -slide- command still worked, sort of. It would attempt to select a particular image off the microfiche, causing the mechanism to shake and make sounds, even illuminate the slide projector’s lamp, which caused things to light up behind the plasma display. Run the code in a loop, and you could get the machine to shake and rumble like a deranged washing machine on spin cycle, along with bright flashes of light as a spaceship exploded. Brendan McGinty, another gamer, vividly remembers arriving on the PLATO scene in August 1974, walking into a room full of PLATO terminals, every single person there playing games, many of them Moonwar, which was also modified to sup
port the -slide- trick. “Every time you shot somebody in Moonwar,” he says, “the slide projector on the old Magnavoxes would go off….These terminals are rockin’ and shakin’ and everything, turnin’ green, and I thought, What the hell is all this stuff? I got hooked on the games.”

  The CERL staff were not so amused. However, in the spirit of Bitzer’s optimistic philosophy, the -slide- hack pointed to what was becoming a steady stream of reasons why the gamers had become beneficial parasites to CERL’s CYBER host: the symbiosis exposed weaknesses, flaws, design oversights, and bugs in the TUTOR language and the system hardware. The -slide- command would in time be fixed. Game authors would have to find other ways to push the system beyond its design limits.

  —

  One PLATO gamer, Silas Warner, didn’t find Moonwar very challenging. “There was never anyone around to play it when you wanted to play it,” he says. He decided to write his own game that was more challenging. In Moonwar, a player’s shot would bounce off the wall (depicted via primitive animation) but it would hit some obstacles. Warner decided to write a new game inspired by Moonwar, but he got rid of the obstacles and just had a square open field. In Warner’s game, instead of players deciding how they were going to move and how they were going to shoot, he thought, why not have a robot player that you could program to move and shoot a certain way? Instead of player versus player, Warner wanted to be able to play against the computer as an option when real players weren’t around, hanging out in the Big Board. His game was called Robotwar, a game that would go on to huge success when reprogrammed to run on the Apple II and sold commercially a few short years later.