Monthly Archives: April 2012

Sentient Software

In 1979 a 30-year-old aspiring science-fiction writer named Mike Berlyn bought an Apple II. He had already finished and delivered his first two novels to Bantam Paperbacks, who would release them under the titles The Crystal Phoenix and The Integrated Man the following year. Now about to start on a third, he had heard that these new PCs were going to change the way writers wrote, and was eager to find out for himself. In the long term, the prediction was of course not wrong, but Berlyn quickly found that the technology of 1979 was, as they say, not quite there yet. The Apple II didn’t even yet support lower-case letters at this point, necessitating all sorts of kludges in early word processors that took them about as far away as you can get from the ideal of what you see is what you get. He ended up writing his third novel, eventually published by Ace Paperbacks as Blight under the pen name Mark Sonders in 1981, the old-fashioned way.

Still, Berlyn was far from disappointed with his purchase. The Apple II may still have been problematic from a practical standpoint, but Berlyn, like so many before and after him, found it an endlessly fascinating toy. When not writing that third book, he spent most of his time exploring his new machine. He found text adventures particularly compelling, but was disappointed by the obvious lack of literary skill of most of the people creating them. Being an enterprising sort, Berlyn decided when the third book was finished that, rather than start right away on a fourth, he’d like to try making a text adventure or two of his own. The result of that aspiration was Sentient Software, a company founded by Berlyn and his wife Muffy with the help of some other partners also located near the Berlyns’ Colorado home. Sentient published two games in 1981, Oo-Topos and Cyborg. Both were written and programmed entirely by Berlyn with a bit of help from his wife, and both were science-fiction adventures involving a damaged spaceship.

In many ways these games are very typical of their era. Technically, they are most similar to Softporn of the games I’ve already discussed on this blog; they are built from a BASIC program with a two-word parser that fetches text and details of the storyworld as needed from data files stored on the disk. They are, in other words, about equivalent to the Scott Adams games in their parser and in the depth of their world modeling, but their use of the disk drive gives them space to be much more loquacious (certainly an important attribute for a “real” writer like Berlyn) and to have much bigger geographies. Indeed, their worlds are quite big ones, but made up mostly of empty rooms, connected via undescribed exits that necessitate painstaking mapping — and that’s outside the obligatory mazes. And of course, the parser makes many puzzles much harder than they ought to be. (Finding out what the correct verbs are, Cyborg tells us, is “half the fun.” Um, no.)

Yet in other ways these games represent something new and significant. Berlyn was the first author to come to the text adventure from the world of traditional fiction. He was interested in the form not, like the hackers who preceded him, as an interesting technical challenge, but rather as a potential new form of storytelling. The packaging of the games emphasized that they were not about “treasures” or “score,” but about “character development,” consistency, and plot. Some of those claims may have been more than a bit of a stretch, but Berlyn was trying, and that is significant in itself.

The plot of Cyborg, the more thematically audacious of the two games, casts you as, well, a cyborg, a human who has been physically and mentally merged with a robot. When play begins, you have amnesia, an adventure-game trope that would soon become a cliché but that may just see its first appearance here. Robbing your avatar of her memory allows Berlyn to place the two of you in the same mental situation. You both spend the game piecing together what brought you to this state, marooned on a stricken spaceship in orbit around a strange planet. Although you are expected to eventually repair the spaceship and lead your people — whom you eventually realize are colonists stored in suspended animation aboard the ship — to the planet below, the vast majority of the plot is not really story per se, but rather backstory, a frame to contain the game’s traditional puzzle- and mapping-oriented play. Within that frame, however, the game’s environments are indeed consistent and believable in a way that hadn’t been seen before. Like amnesia, Cyborg‘s piece-together-the-back-story approach to plotting would soon become an adventure-game cliché. Still, it became a cliché because, at least in these earlier, less jaded days, it worked. Here it allows Berlyn to present a much richer fictional experience than would normally be possible given the primitive technology on-hand to him. His use of it marks him as — and I don’t use this word lightly — a visionary, someone thinking about the medium’s potential in a very progressive way.

One of the most interesting aspects of Cyborg is its handling of the player / avatar split. You play a disembodied human intelligence who must communicate with another, synthetic entity to accomplish absolutely everything. The idea of a split or disembodied consciousness was one that Berlyn found endlessly intriguing; his first two novels both dealt with similar themes, and he would return to it yet again (and most famously) in his next game, Infocom’s Suspended. Here he gets huge mileage out of his concept, including using it to account for the limitations of his parser:


The game’s simple hint system is likewise integrated into the fiction. You can ask your computerized companion what he thinks about locations or items, and occasionally — very occasionally — will get a helpful suggestion.

This unusual concept makes Cyborg one of the few (only?) text adventures ever written in the first-person plural. And again, it’s reflective of some unusually sophisticated thinking about the medium and its possibilities. Scott Adams and others had previously described the player’s avatar as her “puppet,” and at times seemed to give it a separate consciousness, at least if we can judge from the occasional snappy comebacks it gave to nonsensical or dangerous inputs. But no one had previously devised a scenario where even parser frustrations fitted into the scenario so seamlessly. Cyborg marks the first of a long line of games — and almost as many articles in game theory — to explicitly, consciously (ha!) play with the identities of player and avatar. Berlyn even extends the conceit to the verbs permitted. For instance, you cannot LOOK but must SCAN, and an INVENTORY becomes a BODY SCAN.

Given their obviously limited resources, Berlyn and company did the best they could marketing Oo-Topos and Cyborg. For packaging they used a very minimalist cardboard folder, but did commission some nice science-fiction art for the covers.

Still, and as Chuck Benton was discovering at about the same time, it was getting harder for the bedroom hacker without connections to distributors and the like to get his software into stores. Cyborg received an absolutely glowing review in the influential Softalk magazine: “Cyborg introduces the most exciting advances in adventuring since the original Adventure began the whole wonderful thing.” Yet even that wasn’t enough to overcome Sentient’s distributional problems and make the game a success.

Berlyn designed a couple more games for Sentient in 1982, albeit less ambitious arcade-oriented fare, called Gold Rush and Congo. They similarly didn’t make much of an impact. At this point Berlyn and his partners had some sort of falling out which led him to walk away from the company. Over the next couple of years, said partners funded ports of Berlyn’s adventures to the Atari 400 and 800, the IBM PC, and the Commodore 64, before allowing Sentient to fade quietly out of existence. Berlyn, however, was just getting started in interactive fiction, as we’ll see in later posts.

Cyborg is as fascinating conceptually as it can be frustrating to actually play, but it’s well worth a look by any student of the art of interactive fiction. I’ve therefore made the Apple II disk image available for you.

Next time: we’ll take our first tentative steps across the big pond.


Tags: , ,

The Future Was Here: The Commodore Amiga

As has been something of an open secret for quite a while now, I wrote a book. It’s called The Future Was Here: The Commodore Amiga, it’s published by the MIT Press, and now it’s shipping at last.

As the name would imply, my book is a history of the Amiga, a computing platform that pioneered much of the digital world of today. Indeed, my central thesis is that the Amiga represents the world’s first true multimedia personal computer. Much of the book is devoted to working out the implications of that claim.

One thing I wanted to do with the book, as with this blog, was to not neglect the technology in writing technological history. To understand what allowed the Amiga to, say, pioneer the field of desktop video (something that has become so ubiquitous in this era of YouTube that, like “desktop publishing,” the term has ceased to be a useful signifier), one has to understand a bit about its design, even about how the Amiga got its picture to the screen and how this differed from other contemporary computers. So, and while I don’t neglect culture and sociology, I do delve quite deeply into the inner workings of the machine. At the same time, I keep the jargon to a minimum and, when I do indulge, make it a point to explain it carefully beforehand. I thoroughly believe that any patient and interested reader is capable of understanding this stuff if the author just shows a little bit of care, and that’s the assumption that guided me throughout the writing. In other words: no computer science degrees are required. I’m going to go out on a limb here and say that I think many of you who enjoy this blog will also enjoy the book — even if only one chapter deals directly with games. (Hey, at least it’s one of the longest ones…)

Again as I do on this blog, I wanted to encourage active reading, to encourage you to go out and explore some of this technology and art for yourselves. With that in mind, I’ve created a website for the book that hosts a fair amount of content. The book itself can of course be purchased from many fine bookstores, online or brick and mortar.

Oh, and sorry things have been a little quiet with the blog lately. I should have some more stuff for you within a day or three.



My Eamon Problem

Fair warning — this post is going to be a bit meta. It has two purposes. The first is easily dispensed with: to tell you that I’ve revised my earlier posts on the history of Eamon to reflect what I believe to be a more supportable chronology which does not have the system appearing until late 1979. The rest of what follows describes briefly how I came to my conclusions. This is all rather inside baseball, but those of you thinking of growing up to become digital antiquarians yourselves might be interested in this slice of my poor detail-obsessed life.

Traditional histories have given Eamon a release date of 1980, presumably because the first published article about the system, a piece written by Don Brown himself for Recreational Computing, dates from the summer of that year. I initially saw no reason to doubt the traditional chronology. But then I made contact with John Nelson, founder of the National Eamon Users Club. He dropped a bomb on me by saying he had first played Eamon in 1978, and that at that time there were already four additional scenarios available. As the guy who probably did more for Eamon than anyone else, including its creator, Nelson was a hard fellow to doubt. So I wrote those posts based largely on his chronology, even though I never could manage to feel really confident in it. Ever since, those posts have remained the ones I’m least happy about. My dissatisfaction was such that I recently started rummaging through all of the early Eamon disks again, looking for something that would let me pin a definite date onto at least one of them, and thereby begin to build a chronology. As it happened, I found what I was looking for, and that in turn prompted me to revise the earlier articles and write this post. Before I tell you what I found, however, let me first state some of the misgivings that sent me looking in the first place.

The Apple II actually had two versions of the BASIC language. The original machine had in its ROM a very stripped-down version of the language, one that had been put together quickly by Steve Wozniak himself. This version was soon dubbed “Integer BASIC” because it had no support for floating-point (i.e., decimal) numbers, only integers. Because floating-point numbers are very important to certain types of applications, Apple quickly realized the need for a better, more complete implementation of BASIC. They bought one from Microsoft and spent considerable effort customizing it for the Apple II. They dubbed it Applesoft BASIC upon its release in January of 1978. Applesoft was initially not widely used, however, both because its earliest incarnation was quite buggy and because it was housed on tape or disk rather than in ROM, meaning the user had to load it into RAM to use it. With most machines still equipped with only 16 K of memory in these early days, Applesoft, which consumed 10 K by itself, was impractical for most users. It only really caught on from May of 1979, when Apple began shipping the II Plus with Applesoft in ROM; to run an Integer BASIC program on the II Plus, one had to load that language in from disk.

Yet Eamon is written in Applesoft BASIC. And there’s something else: the standard Eamon needs pretty much all of a 48 K Apple II’s memory. (The master disk did originally contain a special, stripped-down version of the program for 32 K machines.) It’s doubtful that it would even be possible to load Applesoft from disk and still have room for Eamon. Even if it was, a 48 K machine would have been a very unusually powerful one for 1978. After the 48 K Apple II Plus began shipping, however, the larger memory quite quickly became an expected standard.

And there’s the text-adventure chronology problem. Scott Adams first released Adventureland and Pirate Adventure during the second half of 1978 for the TRS-80. These games did not appear on the Apple II until early the following year, where they represent the first text adventures available for that platform. To have developed Eamon in 1978, Brown would have had to either: 1) be aware enough of the TRS-80 world that he played Adams’s games and decided to implement a similarly parser-based interface on the Apple II ; 2) have played Crowther and Woods’s Adventure or one of the other games it spawned on a big institutional computer; or 3) have come up with the concept of the text-adventure interface on his own, from scratch. None of these are impossible, but none seems hugely likely either. Depending on when in 1978 Eamon was released, an early Eamon even creates the somewhat earthshaking possibility that it may have been Brown, not Scott Adams, who first brought the text adventure to the microcomputer. Again, this just doesn’t feel right to me.

And then there’s that Recreational Computing article itself. In it Brown writes, “I know of five additional adventure diskettes.” Nelson, on the other hand, believes that “about 20” adventures were available by 1980. He suggested to me that Brown was perhaps referring to adventures that he himself had not written, but it’s very hard for me to read this sense into the paragraph in question. Nelson’s other suggestion, that the article had just lain on the shelf for many months before being printed, seems equally a stretch. If everything else pointed to an earlier chronology, I could accept such reasoning, but in combination with the other questions it becomes a good deal harder.

And then I found what I was looking for. Eamon #3, The Cave of the Mind, was the first not to be written by Brown himself, being from Jim Jacobson and Red Varnum. At the beginning of one of its programs is an REM statement with an actual date: January 30, 1980. This was enough to tip me back over to something much closer to the traditional chronology, with Brown developing the system in the latter half of 1979 in the wake of the Apple II Plus’s release. Sure, it’s possible that the date in the code of Cave represents a revision date rather than a date of completion or release, even though it doesn’t say this. But weighed together with all the other evidence, I feel pretty confident a later date for Eamon is more likely than an earlier.

None of this is meant to criticize John Nelson, who generously shared his memories with me. It’s just that 30 years is a long time. It’s also possible that Nelson might have played an earlier proto-Eamon, presumably written in Integer BASIC for an Apple II with much less memory, which Brown expanded at a later date into the Eamon we know today. Yet unless some real documentary evidence surfaces, or Brown suddenly starts talking, that remains only speculation.

So, the current Eamon articles still represent something of a best guess, and as such I’m still not entirely happy with them. But I think it’s a better guess than the one I made the first time around. Barring more new data, that will have to do.


Tags: ,

Castle Wolfenstein

One night circa early 1981, Silas Warner of Muse Software dropped by a local 7-Eleven store, where he saw an arcade game called Berzerk.

Berzerk essentially played like an interactive version of the programming game Warner had just finished writing on the Apple II, Robot War. The player controlled a “humanoid” who looked more than a little like a robot himself, battling an array of other robots each equipped with their own armaments and personalities. But most impressively, Berzerk talked. The enemy robots shouted out science-fiction cliches like “Intruder alert!” and, Dalek style, single-word imperatives like “Attack!,” “Kill!,” and “Destroy!” Warner was entranced, especially considering that one of Muse’s flagship products was Warner’s own The Voice, an Apple II voice-synthesis system. Still, he’d had enough of robots for a while.

Then one night the old World War II flick The Guns of Navarone came on the television. The most successful film of 1961, it’s the story of a tiny group of Allied commandos who make their way across a (fictional) Greek island to destroy a vital German gun installation. Like most films of its ilk, it can be good escapist fun if you’re in the right frame of mind, even if most of its plot is forehead-slappingly silly. After seeing Navarone, Warner started thinking about whether it might be possible to replace robots with Nazis. One nice thing about filmic Nazis, after all, is that they tend to be as aggressively stupid as videogame robots, marching blithely into trap after ambush after deception while periodically shouting out “Achtung!,” “Jawohl!,” and “Sieg Heil!” in lieu of Berzerk‘s “Attack!,” “Kill!,” and “Destroy!” (One imagines that the Greeks in the movie, when not engaging in ethnically appropriate song and dance or seducing our heroes with their dewy-eyed, heroic-resistance-fighter gazes, must be wondering just how the hell they managed to get themselves conquered by this bunch of clowns.) Other elements of the movie also held potential. The heroes spend much of the latter half disguised in German uniforms, sneaking about until someone figures out the ruse and the killing has to start again. What a game mechanic!

So, from the odd couple of Berzerk and The Guns of Navarone was born Castle Wolfenstein.

Given Wolfenstein‘s position in the history of ludic narrative, it’s appropriate that it should have resulted from the pairing of an arcade game with a work of fiction. Wolfenstein was the first game to unify the two strands of computer gaming I described in my previous post, combining a real story and fictional context with action mechanics best carried out with a joystick or set of paddles. Yet even this gameplay also demanded considerable thought, even strategizing, for success. In the console world, Warren Robinett had attempted a similar fusion a couple of years earlier with the Atari VCS game Adventure, which was directly inspired by Crowther and Woods’s game of the same name. Still, the VCS was horribly suited to the endeavor. Because it couldn’t display text at all, Adventure couldn’t set the scene like Wolfenstein did when the player first started a game. The following is mouthed by a dying cellmate in the castle/fortress in which you are being held prisoner:







Once into the game proper the text dries up, but there are still elements that make it feel like some facsimile of a real situation rather than an exercise in abstract arcade mechanics. The “verbs” available to the player are very limited in comparison to, say, even an old-school text adventure: move, aim, shoot, search a surrendered soldier or corpse, open a door or chest, throw a grenade, use a special item, take inventory. Yet the game’s commitment to simulation is such that this limited suite of actions yields a surprising impression of verisimilitude. One can, for example, use a grenade to blow up guards, but one can also use it to blast holes in walls. Such possibilities make the game a tour de force of early virtual worldbuilding; arguably no one had created a simulated world so believable on such a granular level prior to Wolfenstein.

There is even some scope for moral choice. If you catch them by surprise, guards will sometimes lift their arms in surrender, at which point you are free to kill them or leave them alive, as you will. Similarly, the game allows different approaches to its central problem of escape. One can attempt to methodically dispatch every single guard in every single room, but one can also try to dodge past them or outrun them, only killing as a last resort. Or one can find a uniform, and (in the game’s most obvious homage to The Guns of Navarone) try to just walk right out the front door that way. These qualities have led many to call Wolfenstein the first ancestor of the much later genre of stealth-based games like Metal Gear Solid and Thief. I don’t know as much about such games as I probably ought to, but I see no reason to disagree. The one limiting factor on the “sneaking” strategy is the need to find those battle plans in order to achieve full marks. To do that you have to search the various chests you come across, something which arouses the guards’ suspicion. (These may be videogame Nazis, but they aren’t, alas, quite that stupid.)

In order to make the game a replayable exercise (shades of the arcade again), the castle is randomly stocked with guards and supplies each time the player begins a new game. In addition, play progresses through a series of levels. The first time you play you are a private, and things are appropriately easier — although, it should be noted, never easy; Wolfenstein is, at least for me, a punishingly difficult game. Each time you beat the game on a given level, you increase in rank by one, and everything gets more difficult the next time around. The ultimate achievement is to become a field marshal.

In Warner’s own words, he threw “everything” Muse had on their shelf of technical goodies into Wolfenstein. For instance, we once more see here the high-res character generator Warner had also used in Robot War.

But most impressive was the inclusion of actual speech, a first for a computer game. To really appreciate how remarkable this was, you first have to understand how extraordinarily primitive the Apple II’s sound hardware actually was. The machine contained no sound synthesizer or waveform generator. A program could make sound only by directly toggling current to the speaker itself. Each time it did this, the result was an audible click. Click the speaker at the appropriate frequency, and you could create various beeps and boops, but nothing approaching the subtlety of human speech — or so went the conventional wisdom. The story of Wolfenstein‘s talking Nazis begins back in 1978, when a programmer named Bob Bishop released a pair of programs called Apple-Lis’ner and Appletalker.

Every Apple II shipped with a port that allowed a user to connect to it a standard cassette drive for storage, as well as the internal hardware to convert binary data into sound for recording and vice versa. Indeed, cassettes were the most common storage medium for the first few years of the Apple II’s life. Bishop realized that, thanks to the cassette port, every Apple II effectively contained a built-in audio digitizer, a way of converting sound data into binary data. If he attached a microphone to the cassette port, he should be able to “record” his own speech and store it on the computer. He devised a simplistic 1-bit sampling algorithm: for every sample at which the level of the incoming sound was above a certain threshold, click the speaker once. The result, as played back through Appletalker, was highly distorted but often intelligible speech. Warner refined Bishop’s innovations in 1980 in The Voice. It shipped with a library of pre-sampled phonemes, allowing the user to simply enter text at the keyboard and have the computer speak it — if the program properly deduced what phoneme belonged where, of course.

For Wolfenstein, Warner took advantage of an association that Muse had with a local recording studio, who processed Muse’s cassette software using equalizers and the like to create tapes that Muse claimed were more robust and reliable than those of the competition. Warner: “We went down there [to the studio] one fine day, and I spent several hours on the microphone saying, ‘Achtung!'” Given the primitive technology used to create them (not to mention Warner’s, um, unusual German diction), Wolfenstein‘s assorted shouts were often all but indecipherable. Rather than hurting, however, the distortion somehow added to the nightmare quality of the scenario as a whole, increasing the tension rather than the contrary.

Castle Wolfenstein

Warner’s magnum opus as a designer and programmer, Castle Wolfenstein remained Muse’s most successful product and reliable seller from its release in September of 1981 through Muse’s eventual dissolution, not only in its original Apple II incarnation but also in ports to the Atari 400 and 800, MS-DOS, and (most notably) the Commodore 64. Muse produced a belated sequel in 1984, Beyond Castle Wolfenstein, in which the player must break into Adolf Hitler’s underground bunker to assassinate the Fūhrer himself rather than break out of a generic Nazi fortress. However, while Warner was involved in design discussion for that game, the actual implementation was done by others. The following year, Muse suddenly collapsed, done in by a string of avoidable mistakes in a scenario all too common for the early, hacker-led software publishers. Warner stayed in the games industry for another decade after Muse, but never found quite the creative freedom and that certain spark of something that had led to Robot War and Castle Wolfenstein in his banner year of 1981. He died at the age of 54 in 2004. Wolfenstein itself, of course, lived on when id Software released Wolfenstein 3D, the precursor to the landmark Doom, in 1992.

Whether we choose to call Castle Wolfenstein the first PC action adventure or the first stealth game or something else, its biggest importance for ludic narrative is its injection of narrative elements into a gameplay framework completely divorced from the text adventures and CRPGs that had previously represented the category on computers. As such it stands at the point of origin of a trend that would over years and decades snowball to enormous — some would say ridiculous — proportions. Today stories in games are absolutely everywhere, from big-budget FPSs to casual puzzlers. With its violence and cartoon-like Nazi villains, Wolfenstein is perhaps also a harbinger of how cheap and coarse so many of those stories would be. But then again, we can’t really blame Warner for that, can we?

If you’d like to try Silas Warner’s greatest legacy for yourself, you can download the Apple II disk image and manual from here.

Next time we have some odds and ends to clean up as we begin to wrap up 1981 at last.


Tags: , ,

This Game Is Over

Before the famous Videogame Crash of 1983 there was the Videogame Crash of 1976. By that year Atari’s Pong had been in arcades for four years, along with countless ball-bouncing variants: Handball, Hockey, Pin Pong, Dr. Pong, and of course Breakout. The public was already growing bored of all of them, as well as with the equally simplistic driving and shooting games that made up the rest of arcade fare. As videogame revenues declined, pinball, the form they were supposed to have superseded, started to make a comeback. Even Atari themselves started a pinball division, as manufacturers began applying some of the techniques they’d learned in videogames to a new generation of electromechanical pinball tables that rewarded players with lots of sounds, flashing lights, and high-score leaderboards. When Atari introduced its VCS home-game console in October of 1977, sales were predictably sluggish. Then, exactly one year later, Space Invaders arrived.

Developed by the Japanese company Taito and manufactured and sold in North America under license by Midway, Space Invaders had the perfect theme for a generation of kids entranced with Star Wars and Close Encounters. Its constant, frenetic action and, yes, the violence of its scenario also made it stand out markedly from comparatively placid games like Pong and Breakout. Space Invaders became the exemplar of videogames in general, the first game the general public thought of when one mentioned the form. With coin-operated arcade games suddenly experiencing a dramatic revival, sales of the Atari VCS also began to steadily increase. Thanks to a very good holiday season, sales for 1979 hit 1 million.

However, the real tipping point that would eventually result in Atari VCSs in more than 15% of U.S. homes came when Manny Gerard and Ray Kassar, Atari’s vice president and president respectively, negotiated a deal with their ostensible rivals Taito and Midway to make a version of Space Invaders for the VCS. Kassar is known today as the man who stifled innovation at Atari and mistreated his programmers so badly that the best of them decided to form their own company, Activision. Still, his marketing instinct at this moment was perfect. Kassar predicted that Space Invaders would not only be a huge hit with the VCS’s existing owners, but that it would actually sell consoles to people who wanted to play their arcade favorite at home. He was proven exactly right upon the VCS Space Invaders‘s release in January of 1980. The VCS, dragged along in the wake of the game, doubled its sales in 1980, to 2 million units.

Atari took the lesson of Space Invaders to heart. Instead of investing energy into original games with innocuously descriptive titles like Basketball, Combat, and Air Sea Battle, as they had done for the first few years of the VCS, they now concentrated on licensing all of the big arcade hits. Atari had learned an important lesson: that the quantity and quality of available software is more important to a platform than the technical specifications of the platform itself. This fact would allow the Atari VCS to dominate the console field for years despite being absurdly primitive in comparison to competition like the Intellivision and the Vectrex.

Apple was learning a similar lesson at this time in the wake of the fortuitous decision that Dan Bricklin and Bob Frankston made to first implement VisiCalc on the Apple II. Indeed, one could argue that the survivors from the early PC industry — companies like Apple and, most notably, Microsoft — were the ones that got the supreme importance of software, while those who didn’t — companies like Commodore, Radio Shack’s computer division, and eventually Atari itself — were the ones ultimately destined for the proverbial dustbin of history. Software like VisiCalc provided an answer to the question that had been tripping up computer hobbyists for years when issued from the mouths of wives, girlfriends, and parents: “But what can you really do with it?” A computer that didn’t have a good base of software, no matter how impressive its hardware, wasn’t much use to the vast majority of the public who weren’t interested in writing their own programs.

With all this in mind, let’s talk about computer games (as opposed to console games) again. We can divide entertainment software in these early years into two broad categories, only one of which I’ve so far concerned myself with in this blog. I’ve been writing about the cerebral branch of computer gaming, slow-paced works inspired by the tabletop-gaming and fiction traditions. These are the purest of computer games, in that they existed only on PCs and, indeed, would have been impossible on the game consoles of their day. They depend on a relatively large memory to hold their relatively sophisticated world models (and, increasingly, disk storage to increase the scope of possibility thanks to virtual memory); a keyboard to provide a wide range of input possibilities; and the ability to display text easily on the screen to communicate in relatively nuanced ways with their players.

The other category consists of arcade-style gameplay brought onto the PC. With the exception of the Atari 400 and 800, none of the earliest PCs were terribly suited to this style of game, lacking sprites and other fast-animation technologies and often even appropriate game controllers. Yet with the arcade craze in full bloom, these games became very, very popular. Even the Commodore PET, which lacked any bitmapped graphics mode at all, had a version of Breakout implemented entirely in “text” using the machine’s extended ASCII character set.

On a machine like the Apple II, which did have bitmapped graphics, such games were even more popular. Nasir Gebelli and Bill Budge were the kings of the Apple II action game, and as such were known by virtually every Apple II hobbyist. Even Richard Garriott, programmer of a very different sort of game, was so excited upon receiving that first call from California Pacific about Akalabeth because CP was, as everyone knew, the home of Budge. If Computer Gaming World is to be believed, it was not Zork or Temple of Apshai or Wizardry that was the bestselling Apple II game of all time in mid-1982, but rather K-Razy Shootout, a clone of the arcade game Berzerk. They may have sold in minuscule numbers compared to their console counterparts and may not have always looked or played quite as nicely, but arcade-style games were a big deal on PCs right from the start. When the Commodore VIC-20 arrived, perched as it was in some tenuous place between PC and game console, the trend only accelerated.

You may have noticed a theme in my discussion of these games in this post and in a previous post: many of these games were, um, heavily inspired by popular coin-operated arcade games. In the earliest days, when the PC-software industry was truly minuscule and copyright still a foreign concept to many programmers, many aspired to make unabashed clones of the latest arcade hits, down to the name itself. By 1980, however, this approach was being replaced by something at least a little more subtle, in which programmers duplicated the gameplay but changed the title and (sometimes, to some extent) the presentation. It should be noted that not all PC action-game programmers were cloners; Gebelli and Budge, for instance, generally wrote original games, and perhaps therein lies much of their reputation. Still, clones were more the rule than the exception, and by 1981 the PC software industry had grown enough for Atari to start to notice — and to get pissed off about it. They took out full-page advertisements in many of the big computer magazines announcing “PIRACY: THIS GAME IS OVER.”

Some companies and individuals have copied Atari games in an attempt to reap undeserved profits from games that they did not develop. Atari must protect its investment so that we can continue to invest in new and better games. According, Atari gives warning to both the intentional pirate and to the individuals simply unaware of the copyright laws that Atari registers the audiovisual works associated with its games with the Library of Congress and considers its game proprietary. Atari will protect its rights by vigorously enforcing these copyrights and by taking the appropriate action against unauthorized entities who reproduce or adapt substantial copies of Atari games, regardless of what computer or other apparatus is used in their performance.

In referring to cloning as “piracy,” Atari is conflating two very separate issues, but they aren’t doing so thoughtlessly — there’s a legal strategy at work here.

Literally from the dawn of the PC era, when Bill Gates wrote his famous “Open Letter to Hobbyists,” software piracy was recognized by many in the industry as a major problem, a problem that some even claimed could kill the whole industry before it got properly started. Gates considered his letter necessary because the very concept of commercial software was a new thing, as new as the microcomputer itself. Previously, programs had been included with hardware and support contracts taken out with companies like IBM and DEC, or traded about freely amongst students, hackers, and scientists on the big machines. In fact, it wasn’t at all clear that software even could be copyrighted. The 1909 Copyright Act that was still in effect when Gates wrote his letter in January of 1976 states that to be copyrightable a work must be “fixed in a tangible medium of expression.” One interpretation of this requirement holds that an executable computer program, since it lives only electronically within the computer’s memory, fails the tangibility test. The Copyright Act of 1976, a major amendment, failed to really clarify the situation. Astonishingly, it was only with the Computer Software Copyright Act of 1980 that it was made unambiguously clear that software was copyrightable in the same way as books and movies and that, yes, all those pirates were actually doing something illegal as well as immoral.

But there was still some confusion about exactly what aspect of a computer program was copyrightable. When we’re talking about copyright on a book, we’re obviously concerned with the printed words on the page. When we’re talking about copyright on a film, we’re concerned with the images that the viewer sees unspooling on the screen and the sounds that accompany them. A computer program, however, has both of these aspects. There’s the “literary” side, the code to be run by the computer, which in many cases takes two forms, the source code written by the programmer and the binary code that the computer actually executes after the source has been fed through an assembler or compiler. And then there’s the “filmic” side, the images that the viewer sees on the screen before her and the sounds she hears. The 1980 law defines a computer program as a “set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result.” Thus, it would seem to extend protection to source and executable code, but not to the end experience of the user.

Such protection was not quite enough for Atari. They therefore turned to a court case of 1980, Midway vs. Dirkschneider. Dirkschneider was a small company who essentially did in hardware what many PC programmers were doing in software, stamping out unauthorized clones of games from the big boys like Atari and Midway, then selling them to arcade operators at a substantial discount on the genuine article. When they started making their own version of Galaxian, one of Midway’s most popular games, under the name Galactic Invader, Midway sued them in a Nebraska court. The judge in that case ruled in favor of the plaintiff, on the basis of a new concept that quickly became known as the “ten-foot rule”: “If a reasonable person could not, at ten feet, tell the difference between two competitive products, then there was cause to believe an infringement was occurring.”

So, in conflating pirates who illegally copied and traded software with cloners who merely copied the ideas and appearance of others’ games, implementing them using entirely original code, Atari was attempting to dramatically expand the legal protections afforded to software. The advertisement is also, of course, a masterful piece of rhetoric meant to tar said cloners with the same brush of disrepute used for the pirates, who were criticized in countless hand-wringing editorials in the exact same magazines in which Atari’s advertisement appeared. All of this grandstanding moved out of the magazines and into the courts in late 1981, via the saga of Jawbreaker.

The big arcade hit of 1981 was Pac-Man. In fact, calling Pac-Man merely “big” is considerably underestimating the matter. The game was a full-fledged craze, dwarfing the popularity of even Space Invaders. Recent studies have shown Pac-Man to still be the most recognizable videogame character in the world, which by extension makes Pac-Man easily the most famous videogame ever created. Like Space Invaders, Pac-Man was an import from Japan, created there by Namco and distributed, again like Space Invaders, by Atari’s arch-rival of the standup-arcade world, Midway. Said rivalry did not, however, prevent the companies from working out a deal to get Pac-Man onto the Atari VCS. It was to be released just in time for Christmas 1981, and promised to be the huge VCS hit of the season. Kassar and his cronies rubbed their hands in anticipation, imagining the numbers it would sell — and the number of VCSs it would also move as those who had been resistant so far finally got on the bandwagon.

Yet long before the big release day came, John Harris, Ken Williams’s star Atari 400 and 800 programmer at On-Line Systems, had already written a virtually pixel-perfect clone of the game after obsessively studying it in action at the local arcade. Ken took one look and knew he didn’t dare release it. Even leaving aside Atari’s aggressive attempts to expand the definition of software “piracy,” the Pac-Man character himself was trademarked. Releasing the game as-is risked lawsuits from multiple quarters, all much larger and richer in lawyers than On-Line Systems. The result could very well be the destruction of everything he had built. Yet, the game was just so damn good. After discussing the problem with others, Ken told Harris to go home and redo the game’s graphics to preserve the gameplay but change the theme and appearance. Harris ended up delivering a bizarre tribute to the seemingly antithetical joys of candy and good dental hygiene. Pac-Man became a set of chomping teeth; the dots Life Savers; the ghosts jawbreakers. Every time the player finished a level, an animated toothbrush came out to brush her avatar’s teeth. None of it made a lot of sense, but then the original Pac-Man made if anything even less. Ken put it out there. It actually became On-Line’s second Pac-Man clone; another one called Gobbler was already available for the Apple II.

Meanwhile Atari, just as they had promised in that advertisement, started coming down hard on Pac-Man cloners. They “persuaded” Brøderbund Software to pull Snoggle for the Apple II off the market. They “convinced” a tiny publisher called Stoneware not to even release theirs, despite having already invested money in packaging and advertising. And they started calling Ken.

The situation between On-Line and Atari was more complicated than the others. Jawbreaker ran on Atari’s own 400 and 800 computers rather than the Apple II. On the one hand, this made Atari even more eager to stamp it out of existence, because they themselves had belatedly begun releasing many of their bestselling VCS titles (a group sure to include Pac-Man) in versions for the 400 and 800. On the other hand, though, this represented an opportunity. You see, Harris had naively given away some copies of his game back when it was still an unadulterated Pac-Man. Some of these (shades of Richard Garriott’s experience with California Pacific) had made it all the way to Atari’s headquarters. Thus their goals were twofold: to stamp out Jawbreaker, but also if possible to buy this superb version of Pac-Man to release under their own imprint. Unfortunately, Harris didn’t want to sell it to them. He loved the Atari computers, but he hated the company, famous by this time for their lack of respect for the programmers and engineers who actually built their products. (This lack of respect was such that the entire visionary team that had made the 400 and 800 had left the company by the time the machines made it into stores.)

At the center of all this was Ken, the very picture of a torn man. He wasn’t the sort who accepts being pushed around, and Atari were trying to do just that, threatening him with all kinds of legal hellfire. Yet he also knew that, well, they kind of had a point; if someone did to one of his games what On-Line was doing to Pac-Man, he’d be mad as hell. Whatever the remnants of the hippie lifestyle that hung around On-Line along with the occasional telltale whiff of marijuana smoke, Ken didn’t so much dream of overthrowing the man as joining him, of building On-Line into a publisher to rival Atari. He wasn’t sure he could get there by peddling knockoffs of other people’s designs, no matter how polished they were.

Thanks largely to Ken’s ambivalence, the final outcome of all this was, as tends to happen in real life, somewhat anticlimactic. On-Line defied Atari long enough to get dragged into court for a deposition, at which Atari tried to convince the judge to grant a preliminary injunction forcing On-Line to pull Jawbreaker off the market pending a full trial. The judge applied the legal precedent of the ten-foot rule, and, surprisingly, decided that Jawbreaker looked different enough from Pac-Man to refuse Atari’s motion. You can judge for yourself: below is a screenshot of the original arcade Pac-Man pair with one of Jawbreaker.

Atari’s lawyers were reportedly stunned at the rejection, but still, Ken had no real stomach for this fight. He walked out of the courtroom far from triumphant: “If this opens the door to other programmers ripping off my software, what happened here was a bad thing.” Shortly after, he called Atari to see if they couldn’t work something out to keep Jawbreaker on the market but share the wealth.

Right on schedule, Atari’s own infamously slapdash implementation of Pac-Man appeared just in time for Christmas. It moved well over 7 million units to consumers who didn’t seem to care a bit that the ghosts flickered horribly and the colors were all wrong. The following year, On-Line and Harris developed a version of the now authorized Jawbreaker for the Atari VCS, publishing it through a company called Tigervision. It didn’t sell a fraction of what its inferior predecessor had sold, of course, but it did represent a change in the mentality of Ken and his company. Much of the fun and craziness continued, but they were also becoming a “real” company ready to play with the big boys like Atari — with all the good and bad that entails.

Similar changes were coming to the industry as a whole. Thanks to Atari’s legal muscling, blatant clones of popular arcade games dried up. The industry was now big enough to attract attention from outside its own ranks, with the result that intellectual property was starting to become a big deal. Around this time Edu-Ware got sued for its Space games that were a little bit too inspired by Game Designers’ Workshop’s Traveller tabletop RPG; they replaced them with a new series in the same spirit called Empire. Scott Adams got threatened with a lawsuit of his own over Mission Impossible Adventure, and in response changed the name to Secret Mission.

Indeed, 1981 was the year when the microcomputer industry as a whole went fully and irrevocably professional, as punctuated by soaring sales of VisiCalc and the momentous if belated arrival of IBM on the scene. That’s another story we really have to talk about, but later. Next time, we’ll see how the two broad styles of computer gaming met one another in a single game for the first time.

(My most useful sources in writing this post were an article by Al Tommervik in the January 1982 Softline and Steven Levy’s Hackers.)


Tags: ,