Author’s note: I originally intended for this post to cover adventure games, computer role-playing games, wargames and other simulations, a brief look at the home video game market, and finally the rise of hybrids that fused home video game systems with personal computers. In the grand scheme of the story about personal computers that I am trying to tell, it probably does not make sense to lavish nearly 7,000 words on early adventure games alone, but it’s a topic of personal interest to me and the tale grew in the telling.

Play was central to the formation of personal computer culture. For the early hobbyists who were fascinated by the guts of the machine, the computer was a plaything in and of itself. Many of those who joined the hobby in 1975 or 1976 did so because of games: they had experience with the extensive BASIC game culture that circulated in the time-sharing systems of universities, high schools, and even corporations, and wanted to keep playing at home.

Even after the rise of commercial personal computer software, when the first truly useful applications began appearing, games remained by far the most popular software category (counting by number of titles produced and number of units sold, although not by dollar value). One 1980 catalog of Apple II software, for example, lists 265 titles, of which roughly two-thirds are games, from *Ack-Ack *(an anti-aircraft target shooter)to Wipe Off (a *Breakout *clone). The rest of the catalog comprises demos, educational programs, and a smattering of business software. Whatever they might say about the practical value of the personal computer, buyers had an evident hunger for games.[1]

The Early Games and Their Market

Computer owners got their hands on games in one of three ways. In the early years, the most common means would be simply copying a paper or cassette tape from a friend or colleague, whether with the permission of the original author or not. In the early years, most hobbyists treated game software as a commons to be freely shared, just as it had been in the time-sharing culture through cooperatives like DECUS. This peer-to-peer copying would never entirely go away, despite the commercialization of game software and various schemes by publishers to try to prevent it.

Many magazines and books also published “type-ins,” complete computer programs (almost always written in BASIC) intended to be manually entered at the keyboard (and then saved to tape or disk), and these, too, were most often games. Dave Ahl’s *BASIC Computer Games *(first published in 1973 by Digital Equipment Corporation), a collection of over 100 type-ins, reputedly sold one million copies by 1979. Though type-in publication continued through the 1980s, the inherent limits on the length of such programs (only the most dedicated would tackle a type-in that was more than a few hundred lines long) and their reliance on the universality of BASIC (rather than more performant compiled languages) meant that their significance waned as the sophistication of the game market increased. They could serve as fun demos or educational tools for learning to code, but could not compare to similar games available commercially.[2]

Selection from a state capital guessing game type-in, from the first issue of *Softside *(October 1978). The median type-in was a simplistic game or graphical demo like this.

A selection from a type-in for a much more complex adventure game, published in the September 1980 Softside. This goes on for two-and-a-half more pages, and is about the limit of what was feasible in the type-in format for all but the most steadfast of typers.

Finally, of course, there were the commercial titles offered by software publishers. The game business began in the same way as the personal computer hardware business: with hobby-entrepreneurs selling their creations to fellow hobbyists. In July 1976, for example, D.E. Hipps of Miami, Florida offered a Star Trek game written for MicroSoft’s Altair BASIC for $10 (no one at this stage of the industry paid any attention to niceties such as licensing agreements for the use of the Star Trek name). No common standard data storage standard existed; hobbyists employed a mix of paper teletype tapes, cassette storage, and (for the most extravagant) floppy disks. So Hipps opted to distribute his game as printed source code: a type-in! SCELBI (creators of one of the early, pre-Altair hobby computers), offered another Star Trek variant called Galaxy in the same form. By the late 1970s, the convergence of the industry on a small number of popular storage standards (with CP/M dominant) resolved this problem, and most games were distributed in plastic baggies containing instructions and a cassette or floppy disk.[3]

Contents of a typical computer game package circa 1980. The instructions, command reference, special instruction sheet, and cassette would have come together in a plastic baggie.[Ernst Krogtof, retro365.blog]

It didn’t take long for other entrepreneurs to see a business opportunity in making it easier for software authors to publish their games. It took some time for clear business models and market verticals to emerge. No categorial distinction existed between publishers of games and publishers of utility and business software prior to 1980: Personal Software’s first big hit was MicroChess, followed by VisiCalc, followed by (as we’ll soon see) Zork. Programma International’s founder began as a hoarder of Apple II software, much of it acquired from copies unauthorized by the original author, then turned legitimate to sell those authors’ software instead. Softape tried selling bundles of software by subscription, and then started its own newsletter for subscribers, Softalk.

Some magazines went the other way around: *Softside *magazine (located the next town over from *BYTE’s *Peterborough, New Hampshire headquarters) created The Software Exchange (TSE), while Dave Ahl’s *Creative Computing *set up a label called Sensational Software. Type-ins printed in the magazines became a gateway drug to more convenient (and often more complex and interesting) software available for sale on cassette or diskette.[4]

Figure 21: Creative Computing heavily advertised the Sensational Software brand in the pages of the magazine, as in this July 1980 example describing some of their most popular hits and offering a free catalog of their full offering of 400 titles.

The early personal computer game culture imitated what came before it. The boundary between mini- and microcomputer culture was permeated by thousands who used time-sharing systems at work or school and then went home to a hobby computer. Prior to 1977, a game written for a personal computer was almost invariably based on a game drawn from the other side of that boundary.

Barring a few exceptions (such as the PLATO system available at some universities), users interacted with such computer systems through teletypes or video teletypes that alternated sending and receiving text. So, the resulting games were turn-based, purely textual, and relied on strategy and calculation (or pure luck) to win, not timing and reaction speed. These textual games suited the early hobbyists perfectly, since almost all of their computers also had text-only interfaces, whether mechanical teletypes or video displays like the TV Typewriter.

Other than simple quizzes, demos, and guessing games, popular titles included simulations such as Hammurabi, Civil War and Lunar Lander; statistical recreations of sports contests (baseball, basketball, golf, etc.); or classic games or puzzles from the physical world, like checkers, Yahtzee, and the towers of Hanoi. By far the most popular by far, however, judging by the number of variations published and references in hobby magazines, were descendants of Mike Mayfield’s 1971 Star Trek, a strategic game of galactic war against the Klingons.[5]

An example of a user interaction in a Mike Mayfield-style Star Trek game. Each command entered by the user produces a textual response from the computer. There is no continuous display; the SRS command must be used each time the player wants to see their situation. [The Wargaming Scribe, https://zeitgame.net/archives/1770%5D

Figure 22:

Some early personal computers, however, had video controllers with built-in memory, which allowed for more sophisticated interfaces than the simple back-and-forth exchanges of a teletype. Processor Technology, whose VDM-1 display interface could paint characters at arbitrary points on the screen, sold real-time textual games by Steve Dompier like *Trek-80 *(released in 1976, despite the name). Its interface (including a galactic sector map made of text characters and readouts of the Enterprise’s weapon and shield status) updated in real-time in response to player and (simulated) enemy actions, rather than scrolling by one turn at a time. Cromemco, maker of the Dazzler, an Altair-compatible graphics board, offered the only personal computer games to use pixel graphics prior to the Apple II, starting with a version of the seminal *Spacewar *in early 1977. They followed with a suite of similar games such as Tankwar and Dogfight.[6]

Steve Dompier’s Trek-80 was similar in concept to earlier Trek games, but with alternating commands and responses replaced by a continuously displayed textual interface. [Creative Computing (July/August 1977), 44]

Spacewar on the Dazzler couldn’t match the PDP-1 original, but nothing else at the time sported pixel graphics. [Creative Computing (July/August 1977), 43]

After 1977, when computers with graphical displays became more widely available (especially the full-color Apple II), computer games tapped a new vein of inspiration (and imitation): arcade games. Originally commercialized by Atari and its imitators as standalone arcade cabinets in the early 1970s, then moving into homes by the mid-1970s, these games were typically real-time and focused on action. Relatively cheap and easy-to-make, and relatively disposable to the user (few took more than a few minutes to play a complete game), computer action games proliferated by the hundreds and thousands, many of them direct or near clones of pre-existing arcade or home video games.

By 1980, however, there were major innovations that set personal computer games apart from other game media. In-depth simulations, expansive adventures that took hours to solve, and dungeon crawls teeming with a variety of monsters, treasures, and traps provided immersive experiences that the action-oriented video game consoles, did not, and (given their limited memory and storage capacity) could not provide. Once combined with full-color, bitmapped graphics, these games also surpassed anything previously available on their time-sharing predecessors. The era of imitation was definitively over.

Adventure

For several years of my childhood, for reasons that I no longer recall, our family’s Apple IIe computer, equipped with a green-and-black monochrome monitor, resided in my bedroom. Though much of my autobiographical memory is quite hazy, I can clearly remember each of the Apple II games we owned, each with its own 5 ¼-inch-square floppy disk: Syzygy (a space shooter in the vein of Asteroids), One on One: Dr. J vs. Larry Bird, Winter Games, and *Arcticfox *(a sci-fi tank simulator with wireframe graphics).

But the game that truly captured my imagination, the game whose opening sequence and imagery remain etched (monochromatically) in my mind, was King’s Quest II: Romancing the Throne, a 1985 title by Sierra On-Line. The forty-nine-screen, hand-drawn fairy tale kingdom that you explore in the game (via your avatar, King Graham) felt like a vast world of endless possibility compared to the cramped half-court of One on One, the endlessly repeating monotony of a biathlon course in Winter Games, or the sterile polygonal landscape of Arcticfox’s Antarctica. That open-ended feeling was enhanced by the lure of hidden secrets just out of reach, and a freeform text interface that accepted English commands like “THROW APPLE” (though only a tiny subset of the commands you could imagine would actually work). Despite its limitations and many, many frustrations (at age seven or eight, with no hint book and no Internet walkthroughs, I certainly never came close to completing it), it made me feel that I was truly experiencing an adventure.

One of the colorful environments of King’s Quest II. Limited at the time to a monochrome monitor, I never saw it like this.

The adventure game genre originated in a freely shared, text-driven game created in the time-sharing world. The game, which I will call *Adventure *(it is variously called Colossal Cave Adventure, Colossal Cave, Adventure, or simply ADVENT, after the game’s PDP-10 file name)challenged players to find five treasures within a cave complex by navigating a maze, solving puzzles, and defeating a band of axe-wielding dwarves Its author was Will Crowther, a programmer at Bolt, Beranek and Newman (BBN), where he had written core infrastructural software for ARPANET, the first nationwide computer network.

The BBN team that worked on the Interface Message Processors (IMPs), minicomputers that routed messages across the ARPANET network. This photo was probably taken circa 1969-1970, when the first IMPs were delivered. Crowther is second from the right.

In 1975, Crowther went through a painful divorce. He had always enjoyed playing games with his school age daughters, so he began crafting a game on the company’s DEC PDP-10 to help him stay connected with them. Crowther copied the physical structure of Adventure’s cave directly from a portion of the Mammoth complex in Kentucky. (He had met his wife through caving, and they had explored Mammoth together, so the game was also, in a sense, a means of staying connected to his former, married life.) It is probable (though not certain) that Crowther also drew some inspiration from a popular 1973 time-sharing game called Hunt the Wumpus, which required users to use textual clues to find and kill a Wumpus hidden in a system of caves without falling into a pit. But the conceptual structure of *Adventure *(delving into the earth to find treasure and magical artifacts in the face of devious obstacles and armed foes) came from a new game of pencil, paper, and imagination that Crowther was playing with some of his BBN friends, called Dungeons and Dragons.[7]

In Crowther’s words:

…the caving had stopped, because that had become awkward, so I decided I would fool around and write a program that was a re-creation in fantasy of my caving, and also would be a game for the kids, and perhaps had some aspects of the *Dungeons and Dragons *that I had been playing.[8]

Just as in the later King’s Quest II, the player used simple verb-noun commands (such as “TAKE LAMP”) to interact with the world, but lacking a graphical screen with a visible avatar, he or she also used text commands to move about the world, from one room of the cave to the next (e.g., “SOUTH” or “EAST”). Crowther showed the game off to his D&D buddies and his daughters, then took a new job in California, and forgot about it.[9]

Time-sharing games had once propagated gradually from computer to computer via collectives like the Digital Equipment Computer Users’ Society or colleagues and friends mailing paper tapes to one another. But BBN was on the ARPANET, and Crowther had put his game on a public directory in the BBN computer. From there, someone copied it across the network to a computer in a Stanford lab, where a graduate student, Don Woods, found it in early 1977.

Crowther-Woods *Adventure *as someone might have experienced it in the late 1970s, on a DEC video terminal. [Autopilot, CC BY-SA 3.0]

Fascinated by Crowther’s game, Woods contacted him for the FORTRAN source code, and set about expanding it. He increased the scope by adding more rooms, more puzzles, more foes, and more ways to interact with the world; but he also added the ability to save your progress and a point-tracking system with a final objective: to find fifteen treasures and return them to the starting location. Woods’ larger, more polished version of *Adventure *spread rapidly across the time-sharing world, and became an obsession for some, keeping them at the office past midnight in search of that last treasure. (One of Woods’ additions was a setting to allow admins to disable the game during working hours.)[10]

Adventureland

A frizzy-haired Florida man, Scott Adams, was the first to commercialize a version of *Adventure *for the personal computer. He had first fallen in love with computers on a time-sharing terminal at his Miami high school in the late 1960s. He went on to earn a computer science degree and by the late 1970s, was working as a programmer at telecom manufacturer Stromberg-Carlson. On the side he had become an avid home computer hobbyist, purchasing a Sphere computer in 1975 and then a TRS-80 in 1977. Shortly thereafter he discovered *Adventure *on the company time-sharing system and, like many before and after, could not quit playing until he had beaten it.

Adams decided that it would be an interesting challenge to build something similar for the TRS-80. It would have to be much smaller to fit in the sixteen kilobytes of memory he had available. The Crowther-Woods *Adventure *contained 140 distinct locations and ran to eighty kilobytes of (uncompiled) FORTRAN and fifty-four kilobytes of data for the game text. Adams’ *Adventureland *was considerably smaller, with fewer than thirty-five locations—not necessarily to the detriment of gameplay; for example, the cutting lopped off most of Adventure’s huge and torturous mazes.[11]

Adams’ local TRS-80 buddies were impressed enough with his game that he decided to sell it through both The TRS-80 Software Exchange and Creative Computing, who offered it on cassette for $24.95 and $14.95, respectively, in their January 1979 magazine issues. He followed up with a whole series of games, starting with Pirate Adventure, and ported the games from the TRS-80 to other popular computer platforms. His wife Alexis joined the venture as a business manager and game designer, co-authoring Mystery Fun House and Voodoo Castle.[12]

Scott Adams surrounded by his Adventure series. [Adventure international Microcomputer Software Catalog 2, 5 (1981),4]

The adventure game genre is often criticized for absurd and unfair puzzles, which can be guessed at only through trial-and-error, and tedious mazes or other navigational obfuscations. These early games from circa 1980 are among the worst offenders. In Adventureland, for example, a “very thin” black bear blocks your way, and the only way to get past it is to “yell” at it. Feeding this apparently hungry bear honey will prevent you from completing the game, because the honey is one of the treasures you must collect. You could easily get to state in these games where you have lost the game without knowing it.[13]

Yelling at a bear in Adventureland. [https://www.filfre.net/2011/06/adventureland-part-1]

But these criticisms are retrospective: the contemporary press and the buying public lapped up the Adams’ adventures and all of their imitators. We have to remember that the *appeal *of this genre lay in getting immersed (one might say “lost”) in the game for hours every evening, clawing your way forward towards ultimate triumph for weeks, or even months, on end. In a market full of arcade-like games that offered the convenient but shallow fun of a bag of potato chips, adventure games provided a rich and fulfilling meal for the imagination. As one lover of the genre put it:

Adventure is the product of imagination appealing to imagination. It is not just the puzzle, or the theme, or the nonplayer characters and their personalities. It is a verbal tapestry of interwoven phrases that whisk you away to magical kingdoms of the mind. The computer becomes a tool of reaching that conveys you where it will. You go along eagerly, breathlessly, awaiting what comes next.[14]

The catch was that this delicacy was consumable only once: a solved adventure game was no more interesting to revisit than a solved crossword puzzle. So, they *had *to provide a challenge: no one wanted to pay $24.95 for a game on the way home from work and then breeze through it before bedtime. A game that was very fair would also risk being seen as a waste of money. Despite improvements in design in future years that would banish some of the worst practices of the genre, adventure games remained trapped on the horns of this dilemma.[15]

Zork

The Adams’ “Adventure” line made them wealthy enough to build a faux-castle outside Orlando, and kicked off one of the most popular computer game genres of the 1980s. By late 1980, half-a-dozen other companies were putting out personal computer adventure games, from The Programmer’s Guild to Mad Hatter Software, as well as a version of Crowther-Woods *Adventure *put out by Microsoft. But they are overshadowed in the historical record by a competitor that subsequently dominated both sales of and critical attention to text adventure games. It began at MIT. In the spring of 1977, the Crowther-Woods *Adventure *arrived over the ARPANET at the PDP-10 at the Laboratory for Computer Science (LCS), and sank its claws into its employees. Impressed by what Crowther and Woods had done, but convinced that it could be made even better, a group of LCS staff set out in May 1977 to one-up Adventure.[16]

Dave Lebling, who had already worked on several games (including Maze, the first first-person shooter game), kicked off the project. Lebling played Dungeons and Dragons in the same Cambridge D&D group as Crowther had (though not at the same time), and based the game’s combat system on the tabletop game. Then Marc Blank, Tim Anderson, and Bruce Daniels filled in most of the core structure of the program. They gave it the place-holder name of *Zork *(a term used as an inside-joke expletive at LCS, as in “why won’t this zorking thing work”), which ended up sticking permanently. The game reached its completed state in early 1979, by which point it greatly exceeded the original *Adventure *in scale, with 191 rooms and 211 items, occupying a full megabyte of memory.[17]

Coded in a LISP-descendant called MUDDLE or MDL, *Zork *had an elegant design that encapsulated all the information about the possible interactions of each room and item in a single block of code and data, making it much easier to extend than Adventure. It also had a much richer text interface: both *Adventure *and *AdventureLand *accepted only “verb noun” commands, but *Zork *also allowed for conjunctions and prepositions (for example, “TAKE SWORD AND LANTERN FROM SACK”).Though aping the basic tropes of Adventure (a small overland area leading to an underground treasure-hunt), its more complex architecture allowed for a richer and more clever set of puzzles.[18]

In the spring of 1979, several key staff members of LCS were poised to leave MIT. Their supervisor, Al Vezza, proposed to keep the band together by forming a company. Incorporated in June as Infocom, its new employees and shareholders included Lebling, Blank, and Anderson.

While the various partners mulled what exactly to do with their new business, Blank and a fellow LCS alum, Joel Berez, figured out how to cram *Zork *onto a microcomputer: they cut the number of rooms and items in the game in half and removed all the features of MDL not needed for the game, creating an interpreter for a simpler language they called *Zork *Implementation Language (ZIL). The resulting program occupied just seventy-seven kilobytes. To get this to fit into a microcomputer memory half that size, they had one last trick: a virtual memory system built into the interpreter, to swap chunks of the program on and off the disk as needed (typical floppy disk capacities at the time were over 100 kilobytes, and continued to grow). This meant that *Zork *could *only *run off of a floppy drive (whose rapidly spinning disk could sync to a new data location in a fraction of a second and supply data at fifteen kilobytes per second), never a cassette (which took a minute or more to fully unwind or rewind and supplied data at 300 bits per second). Or, to put it another way, the growing market prevalence of affordable floppy drives made larger personal computer adventure games feasible: it took about twenty minutes to load a Scott Adams adventure game from tape.[19]

In late 1979, Blank and Berez convinced a reluctant Vezza (who wanted to get into business software) to make a microcomputer *Zork *Infocom’s first product. They initially published through Personal Software, co-owned by MIT’s own Dan Fylstra, which had just recently released VisiCalc. But after *VisiCalc’s *smash success, Fylstra no longer wanted to deal in mere games, so Infocom became its own publisher for subsequent games—including *Zork II *and III, built from the remaining unused material from the original PDP-10 Zork.

*Zork *became available in December 1980 and sold 10,000 units in 1981, mostly on the Apple II, despite an eye-watering price of $39.95, at a time when most games cost fifteen to twenty-five dollars. Then, astonishingly, in an industry typically characterized by ephemerality and obsolescence, sales continued to grow, year after year. They peaked in 1984 with over 150,000 copies sold. No doubt Zork’s self-referential humor, its restrained but clever marketing, and the high quality of the game itself (certainly the most well-crafted adventure game to date) all helped to sell the game.[20]

The evolution of Zork’s marketing strategy, from the underground-zine feel of this 1981 ad under Personal Software… [*SoftSide *(April 1981)]

To this more austere and elegant pitch under Infocom in 1982. [Softline (September 1982)]

But many sales also must have arisen from the startling impression given by sitting down in a store (or at a friend’s house) to interact with this remarkable piece of software. Bob Liddil, reviewing *Zork *for *BYTE *magazine, pointed to the fluency of the parser as the element that first pulled him in:

I was eager to test Zork’s biggest selling point, intelligent input (ie: its ability to accept free-form instructions). I typed “OPEN THE BAG AND GET THE LUNCH,” in reference to a brown paper sack inside the house. The computer complied. There was water and food, so I typed “EAT THE LUNCH AND DRINK THE WATER,” to which the computer responded with gratitude for satisfying its hunger and thirst. I was hooked.[21]

The game seemed to *understand *the user and to have an appropriate answer (or a witty retort) ready for everything they might try, from expletives (“FUCK > SUCH LANGUAGE IN A HIGH-CLASS ESTABLISHMENT LIKE THIS!”) to attempts to outwit the command system (“FIND HANDS > WITHIN SIX FEET OF YOUR HEAD, ASSUMING YOU HAVEN’T LEFT THAT SOMEWHERE.”), to questions about the imaginary world in which the game is played (“WHAT IS A ZORKMID? > THE ZORKMID IS THE UNIT OF CURRENCY OF THE GREAT UNDERGROUND EMPIRE.”) Along with *VisiCalc *and WordStar, *Zork *functioned not just as a piece of software that did something, but also as an existence proof (for the owner and for skeptical friends and family) that the microcomputer could be more than merely a toy version of a real computer.[22]

Zork sales finally fell off in the mid-1980s, not because new text adventure games had surpassed it (Infocom continued to rule that particular roost, and *Zork *remained their flagship), but because of the steady improvement in personal computer graphics and the corresponding ascendancy of graphical games over textual ones.

Mystery House

The first graphical adventure game actually appeared several months before Zork: On-Line Systems’ *Mystery House, *created by Ken and Roberta Williams. Unlike Scott Adams and most of the early personal computer hobbyists, Ken Williams got into computers for money, not love. Raised in greater Los Angeles in an unhappy home, he was a driven and impatient young man, and graduated high school at just sixteen. Roberta Heuer, a dreamy young woman whom Williams met through a double date, was impressed enough by his intelligence and ambition to give in to his insistence that they marry in 1972, while they were both still teenagers.

With the expectation of children to come, Ken abandoned his physics program at Cal Poly Pomona for a more immediately lucrative career in data processing. His father-in-law helped him get a loan to attend Control Data Corporation’s training school (the Control Data Institute), and from there he went on to a series of positions working on “big iron” batch-processing systems, constantly bouncing from job to job and home to home in search of better opportunities and a fatter pay check. He and Roberta wanted a bigger house and more creature comforts, but most of all they dreamed of an early retirement to a life out-of-doors, far from the city.[23]

Ken and Roberta as newlyweds in 1972.

The Williamses took no notice of microcomputers until Ken and one of his co-workers, Robert Leff, concocted a way to make money off of them: selling fellow programmers a microcomputer implementation of FORTRAN, one of the most popular data processing languages. Not only could this venture make him and Roberta still richer (always a key consideration), it could free them to finally move away from the traffic and grind of Los Angeles and to live out their dream of rural life. Initially Ken planned to write FORTRAN for the TRS-80, but he redirected his energies to the more capable Apple II after he and Roberta got themselves one for a mutual Christmas present.

Meanwhile, Roberta had gotten hooked on adventure games. Ken had an electromechanical teletype terminal in their home for one of his consulting jobs, and connected it to a computer with the Crowther-Woods *Adventure *available to play. He showed the game off to Roberta. For Ken it was a curiosity, but for Roberta it became an obsession: she would not quit until she had beaten the game, weeks later. Ken brought home a borrowed TRS-80 and cassette tapes for the Scott Adams adventure series, and she flew through those, too. Soon she had an idea for a game of her own: instead of a treasure hunt, it would be a murder mystery; a mix of *Clue *and *Ten Little Indians *set in a creepy old Victorian house.

She insisted that Ken help her create it, and, after putting her off several times, he finally relented. Roberta wanted to add pictures of each room as a way to make this new game better than what came before, taking advantage of the Apple II’s 280×192 pixel high-resolution graphics mode. Because storing dozens of bitmapped images on a floppy disk would be impossible, Ken bought a VersaWriter accessory, a tablet with a movable arm that let Roberta capture the (x, y) position of each line endpoint in her pictures and store them into the computer. He wrote code to re-create the pictures from these coordinates by drawing the lines at runtime.[24]

A circa 1983 ad for the Atari version of the VersaWriter.

Like Crowther and Adams, Ken split the data tables apart from the code that interpreted them. This allowed Roberta to work out all of the information about the rooms in the game, the items they contain, and the actions the player can perform, without needing to write any code. This division of labor between programming and design, quite novel to computer game software, came about from the accident of Roberta’s limited technical skills (she had worked briefly as COBOL programmer, at Ken’s insistence) and Ken’s lack of interest in the game: he was still focused on launching Apple FORTRAN.[25]

Then, while visiting local computer stores to pitch his computer language, Ken demoed an early version of Roberta’s game and everyone in the store gathered around to see it. The owners asked when they could have copies to sell. Ken realized he was backing the wrong horse: it was Roberta’s side project that would make them rich, not FORTRAN. Moreover, rather than give up a cut to a publisher like Programma International, they would take all the revenue for themselves, by publishing the game through the company name he had already registered for his never-to-be-released FORTRAN, On-Line Systems. On top of that, they could make even more money by distributing games into the stores they were already visiting on behalf of other software authors, like Scott Adams’ Florida-based Adventure International. Eventually unable to manage both publishing and distribution, he convinced his former colleague and erstwhile FORTRAN partner, Robert Leff, to buy out the distribution business, which grew into the industry behemoth Softsel.[26]

After a month of development on nights and weekends (Ken’s pace was manic: in his memoir he writes that he always strove to be a “Triple-A” player, and his brother called him a “chronic workaholic”), the Williams’ started selling *Mystery House *in May 1980. It required forty-eight kilobytes of memory, but with chip prices falling continuously, this was not so stringent a requirement as it had been even a year before.

The game’s simplistic “mystery” ends with the player gunning down the de facto murderer: the only living character to be found in a houseful of victims. The puzzles are among the more poorly clued and arbitrary to be found in a genre full of such frustrations. But for adventure-starved gamers of the time it was enchanting: not only could they witness the virtual world which they were navigating, it actually changed in response to their actions (picking up an object, like the taunting note that kicked off the murders, would remove it from the scene). Roberta’s drawings, crude and child-like as they certainly are, gave the game a visual appeal that drew in new buyers, and more than justified its price of $24.95.[27]

The entry room to Mystery House. The green and purple colors are artifacts of how the Apple II high-res graphics mode works. [https://www.filfre.net/2011/10/mystery-house-part-2]

That summer Ken and Roberta were pulling in $30,000 a month and shopping for a house far from Los Angeles, in Coarsegold, California, nestled in the foothills of the Sierra Nevada near Yosemite National Park. On-Line Systems became Sierra On-Line. A few months later a second “High-Res Adventure” followed, *The Wizard and the Princess, *which added visually-stunning color to *Mystery House’s *line drawings: Ken used dithering techniques to make the six colors available in high-res mode appear like twenty-one. Roberta’s *King’s Quest *series, which I encountered on my Apple II, did not begin until 1984. It became Sierra’s best seller: by 1987, the first three installments of the series had sold a combined 500,000 copies, at least according to Sierra’s own marketing.[28]

A scene from The Wizard and the Princess. Nothing like this had been seen on microcomputers before. The dark blue of the man’s shirt is composed of dithered blue and black, and the light blue of his pants from blue and white, etc. The odd colors at the borders between different-colored regions were once again artifacts of the Apple II high-res color system.

It stands out, in a story populated almost entirely with male characters, that two of the earliest adventure game designers (Alexis Adams and Roberta Williams) were women. The scope of Alexis’ contributions aren’t entirely clear, but Roberta was arguably the most successful adventure game designer of all time. There was an appeal in the adventure game genre, which had more in common with a mystery novel or a logic puzzle than an arcade game and typically eschewed violence (the summary execution of Mystery House’s killer notwithstanding), that attracted some women to an otherwise almost entirely masculine industry.[29]

Ken and Roberta in 1989 after winning a Software Publisher’s Association Award for Best Role-Playing or Adventure Game for *King’s Quest IV. *[*Sierra News Magazine *(Summer 1990) 8]

In a world where multiple discovery and parallel invention are the norm, it is also remarkable that all of the games we have discussed (and indeed all the computer adventure games ever made) can trace their ancestry to the Crowther-Woods Adventure. In the meantime, though, many other computer game authorshad drawn inspiration from Dungeons and Dragons, spawning an entirely different genre of computer games, more in tune with D&D’s wargaming roots.

---

[1] Programma International, “Spring 1980 Catalog” (Spring 1980), 3-5 (https://ia903201.us.archive.org/12/items/Programma_Catalog_Spring_1980_for_APPLE_II/Programma_Catalog_Spring_1980_for_APPLE_II.pdf).

[2] J.J. Anderson, “Dave tells Ahl—The History of Creative Computing,”*Creative Computing *(November 1984), 72.

[3] “A Star Trek Product,” BYTE (July 1976), 92; “Scelbi Software,” *BYTE *(July 1976), 17.

[4] Alexander Smith, *They Create Worlds: The Story of the People and Companies That Shaped the Video Game Industry, Vol. I: 1971-1982 *(Boca Raton: CRC Press, 2020), 366-368; Jimmy Maher “Adventureland, Part 2,” The Digital Antiquarian (June 24, 2011) (https://www.filfre.net/2011/06/adventureland-part-2); David H. Ahl, “The First Decade of Personal Computing,” *Creative Computing *(November 1984), 30.

[5] Smith, They Create Worlds, 266-267; David H. Ahl, ed., *101 BASIC Computer Games: Microcomputer Edition *(New York: Workman Publishing, 1977).

[6] The Wargaming Scribe, “The beginning of home computer gaming: the VDM-1 and the SOL-20” (August 16, 2023) (https://zeitgame.net/archives/10450); “Cromemco Dazzler Games” (Mountain View: Cromemco, 1977); Steve North, “Two Space Games (With Graphics!) For Your Home Computer,” *Creative Computing *(July/August 1977) 43-44; “Spacewar Available for the Cromemco Dazzler,” *Cromemco News *(January 1977).

[7] Smith, They Create Worlds, 383-384; Katie Hafner, “Will Crowther Interview,” (March 1994), 1-5 (https://archive.org/details/WillCrowtherInterview/mode/1up). I read into

[8] Quoted in Dale Peterson, *Genesis II: Creation and Recreation with Computers *(Reston: Prentice-Hall, 1983), 188.

[9] Smith, They Create Worlds, 384-385; Dennis G. Jerz, “Somewhere Nearby is Colossal Cave,” (2007), 83 (https://jerz.setonhill.edu/resources/preprint/SNiCC.pdf

[10] Smith, They Create Worlds, 384-385; Jerz, “Somewhere Nearby is Colossal Cave,” 13; Jimmy Maher, “The Completed Adventure, Part 1” The Digital Antiquarian (June 2, 2011) (https://www.filfre.net/2011/06/the-completed-adventure-part-1/); Tracy Kidder, The Soul of A New Machine (New York: Little, Brown, 2000 [1981]), 86-89.

[11] IF Archive Adventure zip (https://unbox.ifarchive.org/?url=/if-archive/games/source/adv350-pdp10.tar.gz);”AdventureLand map (https://www.solutionarchive.com/file/id%2C3/); Maher, “AdventureLand, Part 1,” *The Digital Antiquarian *(June 22, 2011) (https://www.filfre.net/2011/06/adventureland-part-1).

[12] Robert Levering, Michael Katz, and Milton Moskowitz, *The Computer Entrepreneurs: Who’s Making It Big and How in America’s Upstart Industry *(New York: NAL Books, 1984), 114-118; Smith, They Create Worlds, 388.

[13] Jimmy Maher, “Adventureland, Part 1,” The Digital Antiquarian (June 22, 2011) (https://www.filfre.net/2011/06/adventureland-part-1).

[14] Bob Liddil, “On the Road to Adventure,” *BYTE *(December 1980), 170.

[15] The 1990 LucasArts adventure, Loom, for example, though it is an artistic masterpiece, was criticized by reviewers for being too short and too easy. Scorpia, “Scorpion’s View: ‘Conquests of Camelot’ and ‘Loom’,” Computer Gaming World (July-August 1990), 51, 63, Simply making the games larger, with more puzzles was technically infeasible in the early years (we have already seen that *Adventureland *had to be much smaller than Adventure to fit on a microcomputer); later, as the costs of game production went up, it became financially infeasible. There is an expert dissection of the sins of one early adventure game in Jimmy Maher, “The Wizard and the Princess, Part 2,” *The Digital Antiquarian *(October 21, 2011) (https://www.filfre.net/2011/10/the-wizard-and-the-princess-part-2).

[16] Bob Liddil, “On the Road to Adventure,” *BYTE *(December 1980), 162.

[17] Jimmy Maher, “The Roots of Infocom,” *Digital Antiquarian *(January 1, 2012) (https://www.filfre.net/2012/01/the-roots-of-infocom); Jimmy Maher, “Zork on the PDP-10,” *Digital Antiquarian *(January 3, 2012) (https://www.filfre.net/2012/01/zork-on-the-pdp-10); Stephen Granade and Philip Jong, “David Lebling Interview,” *Brass Lantern *(undated, ca. 2000) (http://brasslantern.org/community/interviews/lebling.html); Nick Montfort, *Twisty Little Passage: An Approach to Interactive Fiction *(Cambridge: MIT Press, 2003), 86. Eric Roberts, Crowther and Lebling’s dungeon master, ran a variant of D&D he called Mirkwood Tales. Jon Peterson, *Playing at the World *(San Diego: Unreason Press, 2012), 617-618, 622.

[18] P. David Lebling, “*Zork *and the Future of Computerized Fantasy Simulations,” *BYTE *(December 1980), 172-182.

[19] Jimmy Maher, “ZIL and the Z-Machine,” *The Digital Antiquarian *(https://www.filfre.net/2012/01/zil-and-the-z-machine); Maya Posch, “Zork And The Z-Machine: Bringing The Mainframe To 8-bit Home Computers,” *Hackaday *(May 22, 2019) (https://hackaday.com/2019/05/22/zork-and-the-z-machine-bringing-the-mainframe-to-8-bit-home-computers); Scott Adams “Pirate’s Adventure,” *BYTE *(December 1980), 212. Virtual memory was a well-established technique in minicomputer and mainframe operating systems, but no widely used personal computer OS offered virtual memory until the release of Windows 3.0 in 1990.

[20] “InfoCom Shipments By Title and Year” (https://www.flickr.com/photos/textfiles/2419969220); Bob Liddil, “Zork, The Great Underground Empire,” *Byte *(February 1981), 262.

[21] Liddil, “Zork, The Great Underground Empire,” 262.

[22] Jimmy Maher, “Parser Games,” *Digital Antiquarian *(January 16, 2012) (https://www.filfre.net/2012/01/parser-games).

[23] Levy, Hackers, 293-297, 302-303; Ken Williams, *Not All Fairy Tales Have Happy Endings: The Rise and Fall of Sierra On-Line *(Ken Williams, 2020), 12-24, 22-24; Jimmy Maher, “Ken and Roberta,” *The Digital Antiquarian *(October 2, 2011) (https://www.filfre.net/2011/10/ken-and-roberta).

[24] Williams, Not All Fairy Tales, 55-56, 66-68, 88; Levy, Hackers, 303-304; Ken Wiliams, “Introduction to The Roberta Williams Anthology” (1996) (https://wiki.sierrahelp.com/index.php/Introduction_to_The_Roberta_Williams_Anthology). The account in the previous paragraphs is interpolated from the above sources, which are partially contradictory. All differ about who got the Apple II and why. Levy never mentions the TRS-80 or any adventure games besides Adventure, and has Roberta finishing that game after the time the Apple II was purchased, implying she never played any other adventure games before deciding to write Mystery House: the timeline would simply be too tight. I believe this is wrong, and either an intentional elision or a false interpolation by Levy. It is unlikely that the Williamses would later entirely hallucinate having brought home and played the whole series of Scott Adams games. The accounts also differ on whose idea it was to add pictures to the game. I’m inclined to believe it was Roberta, to whom the game idea and all the passion for it belonged.

[25] Williams, Not All Fairy Tales, 69-73.

[26] Levy, Hackers, 308-310; Williams, Not All Fairy Tales, 73; Ken Williams, “A Message From the President,” Sierra News Magazine (Summer 1990), 35].

[27] John Williams, “Sierra’s First Ten Years,” Sierra News Magazine (Summer 1990), 6; Williams, Not All Fairy Tales, 79; Jimmy Maher, “Mystery House, Part 2,” *Digital Antiquarian *(October 9, 2011) (https://www.filfre.net/2011/10/mystery-house-part-2); “Game 57: Mystery House,” Data-Driven Gamer (April 22, 2019) (https://datadrivengamer.blogspot.com/2019/04/game-57-mystery-house.html).

[28] Jimmy Maher, “The Wizard and the Princess, Part 1,” *Digital Antiquarian *(October 20, 2011) (https://www.filfre.net/2011/10/the-wizard-and-the-princess-part-1); “On-Line Systems Presents: Hi-Res Adventure,” Softline (September 1981), 16; Levy, Hackers, 310-311; “Sales Data,” *King’s Quest Omnipedia *(https://kingsquest.fandom.com/wiki/Sales_data#King’s_Quest_Original).

[29] In later years, Sierra On-Line would employ several more women as designers—Lori Cole (the Quest for Glory series), Christy Marx (Conquests of Camelot and Conquests of the Longbow\