<<- Chapter 3   Table of Contents   Chapter 5 ->>

Chapter 4: Infocom


We believe there is a great future for this type of game, both for the 

players and for the implementers and designers… 

-- David Lebling, Marc Blank, and Tim Anderson (1979) 


The Mainframe Zork 

            Like those of virtually all universities, the MIT computer science department was swept up in the Adventure craze of early 1977.  The game was solved by that May through a committed group effort, and a group of hackers promptly began looking into creating their own derivation.  Unlike most, though, this group did not content themselves with merely improving Adventure.  They wanted to create their own piece of IF, one that would be superior in every way to its predecessor.  In this, they were motivated not only by the challenge of the task itself, but also by a certain sense of collegiate pride.  As the two premier computer science schools in the country, MIT and Stanford enjoyed a fierce rivalry with one another.  If Don Woods at Stanford had created the first true adventure game, the hackers at MIT could still create the best.  From this thought process would spring Zork. 


            The group of students and computer lab hangers-on responsible for Zork were a motley crew but an accomplished one.  As described by Tim Anderson, one of their number: 


Marc Blank was enjoying a respite from medical school; I had just finished my master's degree; Bruce Daniels was getting bored with his Ph.D. topic; and Dave Lebling was heartily sick of Morse code (Anderson, “First”). 


They called their game Zork for lack of any better title, the word being one of several nonsense interjections that floated about the MIT computer labs.  An early, primitive version of the game was already on the MIT computers by the end of June, 1977.  Its four core implementers, along with a cast of others who came and went, spent much of their spare time for almost the next two years expanding, tweaking, improving, and debugging the game, while an army of students waited eagerly to explore each new area that was added and, of course, hackers being hackers, did their best to “break” the game.  The final significant addition was made to the game in February of 1979, although bug-swatting would actually continue into 1981.  Development was stopped not by the game being “finished” in any objective sense, but rather by its having grown so huge that there was no memory space left for it on even the relatively powerful DEC mainframes on which it ran.  By this time, the game had spread all over the ARPANET under two different names.  In addition to the original Zork, a FORTRAN port that had been renamed to Dungeon was also making the rounds. 


            By any technical measure, Zork succeeded splendidly in its goal of showing up Adventure.  It was several times as large, its sprawling geography containing literally hundreds of individual rooms.  It was filled with complexities of which Adventure had never dreamt, such as a little boat the player could board to sail about the game’s waterways and some remarkably sophisticated fellow dungeon inhabitants who moved about the game tending to their own business when they were not harrying the hapless player in some way.  By far its most impressive and important trait, though, was its parser.  With the fruits of many years of natural language research at MIT to draw upon, Zork could understand a remarkably sophisticated range of inputs and had a vocabulary of 908 words.  Gone were the days of wrestling with an obstinate two-word parser.  Almost any reasonable input, and a fair number of unreasonable ones, could now be understood.  Virtually all of the conventions of modern IF parsers appear, all at once, with this game, and the general technical paradigm of IF development that languages like Inform still follow today is present and accounted for.  In the sheer sophistication of its programming and world model, Zork represents a quantum leap from anything that existed before it. 


Unfortunately, it is less successful as a game.  Its plot, such as it is, is another iteration of the tried and true “explore the fantasy landscape and collect the treasures,” for no other reason than collecting treasures scores points and thus must be a good thing.  The sheer size of the game, which looks so impressive on a technical specification sheet, makes it nearly unplayable in reality, especially as virtually none of its hundreds of rooms seems to line up logically with its neighbors.  The whole game is effectively a giant maze, one that requires hours and hours of careful mapping from the player before she can even begin to think of solving any actual puzzles or collecting any of the aforementioned treasures.  As modern IF author Robb Sherwin once asserted, “Zork hates its player.”  Paul O’Brian recently returned to the Zork variation Dungeon, and described what he found there: 


Dungeon set itself up unambiguously as the player's antagonist, and it wasn't particularly concerned with telling a story, nor even with describing a world. Plot is nonexistent, and fabulous treasures are described with perfunctory lines like "You see nothing special about the sapphire bracelet." Instead, Dungeon puts its energies into confusing and confounding the player, and wacky map connections are but the tip of the iceberg. Along with the aforementioned mazes, there's the light source, which always runs out at the worst possible times. There's the Round Room, guaranteed to tangle any map. There are the "secret word" puzzles, some of which still perplex me to this day, even though I know how they operate. And of course, there's the thief, whose annoyances are both numerous and legendary. Dungeon wants nothing more than to see you fail, and it's not overly concerned with how much fun you might be having (O’Brian). 


Zork’s ridiculous size and sadistic design actually make it a much less playable game than Adventure.  Its aesthetic qualities are also vastly inferior.  For all of Wood’s extrapolations and enhancements, Adventureremained at its root a simulation of the real-world geography around Mammoth Cave, and this gave it a certain logic and consistency that is totally lacking in Zork.  Here, the geography sprawls wildly in ways that would be impossible in the real world.  In Zork, a modern, industrialized dam sits just a few rooms away from the classical Land of the Dead, which is itself not far away from a troll who has stepped straight out of Tolkien.  Thanks to the game’s method of creation, in which everyone sort of wildly piled in anything that seemed “cool” at the time, it possesses no internal coherency whatsoever.  Throw in a crazy mix of writing styles and the blender effect just gets worse. 


            The hackers at MIT had thus created an excellent game engine, but a less than stellar game.  An article published by Zork’s principle implementers shortly after the game’s completion would seem to indicate that they were already well aware of their game’s shortcomings.  They are clearly thinking of the future here, and of the potential for IF beyond cliched fantasy dungeon crawls: 


Another similar direction would be to change the milieu of the game. Zork, Adventure, and Haunt (the CFS games known to the authors) all flow back to D&D and the literary tradition of fantasy exemplified by J. R. R. Tolkien, Robert E. Howard, and Fritz Leiber. There are, however, other milieus; science fiction is one that comes to mind quickly, but there are undoubtedly others (Lebling, IEEE). 


 In the same article, they also consider the “introduction of magic spells,” which is very prophetic in that the eventual Infocom’s most artistically successful fantasy games would revolve around just that. 


The Birth Of Infocom and The Zork Trilogy

            By May of 1979, shortly after the publication of the above article, most of the implementers of Zork were beginning to drift away from the MIT computer lab.  Real life was catching up with them at last, and many were considering what prospects the future might hold outside the hallowed walls of academia.  One thing seemed clear to everyone: in the future, computing would no longer revolve around the large institutional mainframes that housed Zork.  The home computer revolution was in full swing, and all of the real business opportunity in the field lay with these tiny, primitive machines.  A handful of MIT stalwarts led by Al Vezza, assistant administrator of the university's Lab for Computer Science, decided to form a company to explore this new market.  Among them were the principal creators of Zork.  The name they chose, Infocom, was of a piece with the many futuristic-sounding trademarks that were popping up at the time: Microsoft, Compuware, VisiCorp, etc.  Infocom was not founded with the intention of revolutionizing IF and making of it an art-form.  That would happen accidentally.  The group that founded Infocom rather envisioned creating something for the world of business software, possibly involving some research several had been involved in at MIT involving relational database systems.  Creating such a product, or any other substantial piece of business software, would however take a great deal of time and money.  Infocom needed something it could release quickly to get itself off the ground and generate some cash flow.  Their minds soon turned back to that erstwhile hobby of theirs, Zork. 


            If Infocom could bring Zork to the microcomputer in anything like its original form, it would have a product that would easily trump the likes of the Adventure International games that currently dominated the market.  By this time, a second generation of microcomputers was beginning to hit the market that was considerably more powerful than its predecessors.  The Infocom team concluded that, through some ingenious hacking, they could fit about half of the original mainframe Zork into one of these machines.  A key factor here was that, unlike Scott Adams and his peers, Infocom would not do their actual development on the microcomputers on which their games would eventually be deployed.  They would rather rent time on a DEC mainframe for this purpose, then transfer their finished work to their games’ microcomputer hosts.  Writing a complicated piece of software requires much more computing power than merely running it.  Offloading the development to a mainframe thus allowed Infocom to create games of great sophistication compared to their peers, for, while writing something like Zork on a microcomputer of the day was completely out of the question, running it was not. 


The key to their design was an imaginary computer chip called the "Z-machine." This chip would be able to run Zork (or at least part of it) if the program were coded in a special, very compact language. Then the design called for each personal computer to have a program (called a Z-machine Interpreter Program or ZIP) that would interpret the special Z-machine language and make the computer act the same way that a real Z-machine would. In order to get Zork written in this special language, another language was invented, called Zork Implementation Language (ZIL), similar in many ways to MDL. Marc built a two-stage translator program that would translate a ZIL program, first into an assembly language and then further into the Z-machine language. He also built a ZIP so that a DECSystem-20 could emulate the Z-Machine (Anderson, "Final"). 


The concept of an IF virtual machine had already been pioneered by Scott Adams.  Infocom, however, took the idea to new heights.  Their virtual machine would be a little marvel of clever design, remarkable in its power, compactness, and expandability.  This Z-Machine that Infocom founders Joel Berez and Marc Blank invented during the second half of 1979 is in fact the same that is used, in an updated form, to run much of the IF that is produced today.  Its remarkable longevity is truly a tribute to its designers.  In creating it, Infocom cemented one of modern IF’s most cherished properties: the ability to run a single game file on a wide variety of computing devices.  The Z-Machine today allows IF games, including Zork and the many Infocom classics that would follow it, to be played on devices that their designers never dreamed of, and has allowed them to live on long after the computers of their own time have passed into history. 


            Another key to bringing Zork home was Infocom’s pioneering use of virtual memory.  The Z-Code story file for Zork I occupies 90K of storage alone.  In addition to this, any computer hoping to run the game must also have memory for the actual Z-Code interpreter as well as space for a stack, in-game variables, etc.  The TRS-80 microcomputer, a typical target machine for Infocom in those early days, had just 32K of memory.  Infocom solved this seemingly insolvable problem by making use of the random-access capabilities of the floppy disks that were quickly replacing cassette tapes as the secondary storage system of choice in the microcomputer industry, paging into memory only those parts of the game that were needed at any particular point.  On a tiny machine like the TRS-80, the results could be excruciatingly slow, as the disk drive ground for several seconds after virtually every input, but the fact that the game worked at all was something of a minor miracle.  Now that responses from IF games are virtually instantaneous, players sometimes speak with nostalgia of those times.  One could tell that one had solved a puzzle or otherwise accomplished something significant even before seeing the game’s response when the disk drive ground for a long time following an input, and the sense of anticipation of seeing the game’s eventual response was delicious. 


            Even with such magic, Infocom was forced to do considerable cutting and reorganizing to that first half of Zork to make it fit on a microcomputer.  However, this actually improved the game: 


In the process of being converted from MDL into ZIL, the program became "cleaner" and friendlier. The geographies of the maze and the coal mine were simplified so that the connections were less arbitrary, and in other places complexity was removed whenever it didn't serve a justifiable purpose. For example, there was originally a barrel sitting near the top of Aragain Falls, but it was just a red herring; its only purpose was to lure unsuspecting adventurers inside and carry them over the falls to destruction. The Rainbow Room had its name changed to On the Rainbow, and that meant removing the silly joke about Rockefeller Center and the NBC Commissary. Since the Land of the (Living) Dead (the word "Living" was removed in order to fit the name on the status line) no longer led to the stairway where Zork III later began, the crystal skull (a brand-new treasure) was put there instead (Anderson, “Final”). 


The resulting game was still far from a literary masterpiece, and its design was still rather disjointed, but the product as a whole was worlds better than anything else available at the time.  Reviews, and subsequently sales, reflected this.  Byte Magazine’s gushing response was fairly typical: “No single advance in the science of adventure has been as bold and exciting as the introduction of Zork: The Great Underground Empire” (Liddil). A transcript of interaction with Zork I on the microcomputer will perhaps give some of its flavor, and convey just how advanced its parser and world model were compared with even the mainframe Adventure, much less the likes of Scott Adams’ games: 

Zork: The Great Underground Empire part 1, release 1 


(c) Copyright 1980 Infocom, Inc. All rights reserved. Zork is a trademark of Infocom, Inc. 


West of House 

   You are standing in an open field west of a white house, with a boarded front door. 

   A rubber mat saying 'Welcome to Zork!' lies by the door. 

   There is a small mailbox here. 


   Opening the mailbox reveals a leaflet. 



   Welcome to Zork: The Great Underground Empire 

   Zork: The Great Underground Empire is a game of adventure, danger, and low cunning. In it you will explore some of the most amazing territory ever seen by mortals. 

   No computer should be without one! 

   The original ZORK was created by Tim Anderson, Marc Blank, Bruce 

Daniels, and Dave Lebling. It was inspired by the ADVENTURE game of 

Crowther and Woods. This version was created by Marc Blank, Dave Lebling, Joel Berez, and Scott Cutler. 


North of House 

   You are facing the north side of a white house. There is no door here, and all the windows are boarded up. To the north a narrow path winds through the trees. 


Behind House 

   You are behind the white house. A path leads into the forest to the 

east. In one corner of the house there is a small window which is slightly ajar. 


   You can see what appears to be a kitchen. 


   The kitchen window is closed. 


   With great effort, you open the window far enough to allow entry. 



   You are in the kitchen of the white house. A table seems to have been used recently for the preparation of food. A passage leads to the west and a dark staircase can be seen leading upward. A dark chimney leads down and to the east is a small window which is open. 

   On the table is an elongated brown sack, smelling of hot peppers. A 

bottle is sitting on the table. 

   The glass bottle contains: 

A quantity of water 


pick up what? 


kitchen table: An interesting idea... 

glass bottle: Taken. 

quantity of water: I can't reach that. 


Living Room 

   You are in the living room. There is a door to the east, a wooden door with strange gothic lettering to the west, which appears to be nailed shut, and a large oriental rug in the center of the room. 

   There is a trophy case here. A battery-powered brass lantern is on the trophy case. Above the trophy case hangs an elvish sword of great 



Infocom’s slimmed-down microcomputer Zork was first published by Personal Software in a version for the Tandy TRS-80 in December, 1980, followed shortly by a version for the Apple II.  By late 1981, Personal Software had sold almost 10,000 copies of the two versions combined, not a bad run for a game in those days. 



Personal Software Zork’s rather incongruous cover art 

(Violence actually hardly existed in the game) 


Even though the game did fairly well, games were not really Personal Software's priority.  It had recently published the first usable microcomputer spreadsheet program, VisiCalc, which was selling at an even better clip than Zork, and for a substantially higher price.  Personal Software soon changed its name VisiCorp to capitalize on VisiCalc's success, and elected to pass on further offerings from Infocom.  When Zork II, consisting of parts of the mainframe game that had been left out of Zork I along with a considerable amount of new material from David Lebling, arrived in late 1981, it appeared under only the Infocom name.  Zork II is quite similar in feel to its predecessor, but is less disjointed and uneven and generally more consistent in its writing style and geography.  Its highlight is a charming set-piece that pays tribute to Lewis Carroll’s Alice in Wonderland with a pair of pills that when nibbled cause the player to shrink or grow. 


Even more impressive is the final game of the original Zork TrilogyZork III consists mostly of new material from Marc Blank, with only part of the end-game and one other set-piece puzzle borrowed from the mainframe game.  Its feel is dramatically different.  In place of the wacky, self-referential humor of Zork I and Zork II is an ominous, brooding tone, which is established right from the opening text: 

As in a dream, you see yourself tumbling down a great, dark staircase. All about you are shadowy images of struggles against fierce opponents and diabolical traps. These give way to another round of images: of imposing stone figures, a cool, clear lake, and, now, of an old, yet oddly youthful man. He turns toward you slowly, his long, silver hair dancing about him in a fresh breeze. "You have reached the final test, my friend! You are proved clever and powerful, but this is not yet enough! Seek me when you feel yourself worthy!" The dream dissolves around you as his last words echo through the void.... 


ZORK III: The Dungeon Master 

Copyright 1982 by Infocom, Inc. All rights reserved. 

ZORK is a trademark of Infocom, Inc. 

Release 17 / Serial number 840727 


Endless Stair 

You are at the bottom of a seemingly endless stair, winding its way upward beyond your vision. An eerie light, coming from all around you, casts strange shadows on the walls. To the south is a dark and winding trail. 

Your old friend, the brass lantern, is at your feet. 


The silly treasure hunts of the earlier games are gone, replaced by a quest to find and defeat a mythical Dungeon Master.  When the player finds her way to the end – assuming she does, for this is a very tough game that is nevertheless scrupulously fair in a way its predecessors were not – she will experience events that subvert and re-contextualize not only this game but the whole trilogy.  Zork III possesses a thematic unity and literary quality new to IF. 


The Mature Infocom 


            A five-year golden age began for Infocom with the completion of the Zork Trilogy, during which the company produced some thirty works of IF of unprecedented sophistication, innovation, and consistency.  Certainly, not every game is equal, but the fact that Infocom’s failures are often more interesting than their successes is a tribute to their restless prodding at the boundaries of what constitutes a text adventure.  A generation of players learned to see the Infocom label as a guarantee of quality, promising that they would at the very least have a good time with a solid, bug-free adventure game, and that there was at least a chance of finding within a work of real innovation and artistic merit.  For those of us in the modern IF community, the Infocom canon is the touchstone to which we return again and again.  Whole genres of modern IF have sprung from single titles of Infocom.  That said, it is I think important not to exaggerate Infocom’s admittedly impressive achievements.  For all its creativity, the company was ultimately bound to the whims of a capricious computer game marketplace that was not necessary looking for innovative literary works, and it was also limited by the primitive hardware of its era.  Freed of both these constraints, the modern IF community has created works that arguably surpass anything that came out of Infocom.  Nevertheless, one cannot study IF without studying Infocom, for its work set the standard for quality and many of the ground rules that we still follow today. 


            An examination of the reasons behind Infocom’s remarkable artistic and, for a time, commercial success must begin with the technology employed in its games’ creation.  As mentioned in the previous chapter, the implementers’ system of developing their games on a DEC mainframe and moving only the completed product to microcomputers enabled them to produce games of much greater complexity than competitors who were bound throughout the development process to the tiny micros on which their games would eventually be sold.  On their mainframe, Infocom’s authors wrote in a programming language known as ZIL specifically designed for the production of IF games, having realized that the task was so specialized that a customized language was far superior a choice for the application than a more general purpose one.  Marc Blank: 


I think that it's the whole of it -- and it's all made possible by the language we use. I mean, just imagine if the language that you use, instead of having FORs and NEXTs (BASIC keywords), had operations that dealt with things like people and objects. And it handled that in a very high-level sense, so you didn't have to spend all your time chasing bits around, but could deal in high-level abstractions. And dealing in those terms is much easier, I mean, there's so many things going on in a game like Deadline or Witness or actually any of the new games, Planetfall and Suspended, that you couldn't keep track of them all unless you had a high-level language that was doing it all for you. It would take three to four years to write Deadline in machine language, and then you'd only have it running on one computer (Addams). 


Infocom’s embrace of specialized tools for the creation of IF has been taken as received wisdom by the modern IF community.  Virtually all quality games of the last ten years have been written in IF-specific languages like Inform or TADS, and an announcement of a new game that was written in a general-purpose language is always greeted with a decided skepticism that is borne of long experience.  Likewise, Infocom’s use of a virtual machine – the Z-Machine – and an interpreter for the deployment of its stories remains the standard model today.  In Infocom’s case, the Z-Machine actually brought with it a huge commercial advantage, for it allowed the company to easily sell its game on a wide variety of computing platforms with very little expenditure of time or energy.  A single Infocom title was sometimes available in versions for more than two dozen separate machines that ranged from DEC mainframes to inexpensive Toy ‘R’ Us-style eight-bits.  


            From the consumer’s standpoint, unaware as they were of the technical underpinnings behind Infocom’s work, the company’s most obvious advantage over its competitors was its remarkably sophisticated parser.  No one else in the industry really rivaled Infocom here, despite the claims of the occasional blustering upstart.  Infocom constantly improved its parsing technology throughout the company’s lifetime, adding more shortcuts and convenience features, more vocabulary, more acceptable sentence structures, and even retro-fitting these changes back to its older games when possible.  Even in its early work, “guess the verb” issues rarely cropped up for the average player.  It was not until the late nineties, as Inform and TADS reached their maturity, that the IF world saw parsers equal to Infocom’s.  Modern parsers have finally made strides beyond Infocom’s, but one must remember that Infocom’s games ran on laughably simplistic hardware compared to what we have available today.  When one considers that fact, Infocom’s parser still seems as remarkable an achievement as ever. 


Infocom’s innovative and powerful development system was evidence of its one of its greatest intangible strengths.  Unlike the vast majority of their peers in the early computer games industry, the implementers at Infocom were not scruffy, self-taught bedroom coders.  They were well-grounded in  programming theory and more than capable of developing robust, long-term solutions to their needs.  Although most were not actually engineering or computer science majors, their days spent hacking away in MIT's computer science laboratory now served them well. 


            For all their wizardry, there was one area of computer technology that Infocom dismissed entirely for many years: only in their dying gasps did they begin including graphics in their game.  When asked about this choice, Infocom usually invoked the somewhat clichéd truism of readers that the best images are those generated by the imagination, and in fact its marketing materials attempted to make its games’ lack of multimedia flash into a selling point: 


We stick our graphics where the sun don’t shine.  You’ll never see Infocom’s graphics on any computer screen.  Because there’s never been a computer built by man that could handle the images we produce.  We draw our graphics from the limitless imagery of your imagination – a technology so powerful, it makes any picture that’s ever come out of a screen look like graffiti by comparison.  And nobody knows how to unleash your imagination like Infocom.  Through our prose, your imagination makes you part of our stories… 


In a way, Infocom’s disavowal of graphics was a very brave choice.  By 1982, a very significant number of its competitors were following the path blazed by Sierra’s Mystery House and including images of some sort in their IF games, and a few years later Infocom stood virtually alone in the marketplace in placing its faith in pure text.  Graham Nelson has estimated that roughly two-thirds of the IF games released during the commercial era were illustrated, a fact that is often forgotten today (Nelson 351).  Some might argue that by sticking to its guns on this issue Infocom speeded its downfall, although there were enough other factors involved there that this is by no means an open and shut case to make.  What does seem inarguable is that  devoting all of the resources of the limited computer technology of the era to prose and  world modeling rather than graphical displays allowed Infocom to create games of much greater complexity than would otherwise have been possible.  There is something else, too, which bears directly on Infocom’s current hallowed reputation in the modern IF community.  The graphics in its competitor’s contemporaneous games are so poor by modern standards that they distract the player from whatever virtues the games might otherwise possess.  Infocom’s games, because they do not rely on now-dated visual flash, have a timeless quality about them that their competitors’ titles lack. 


            Infocom insisted fanatically on quality control.    The company’s testing process became something of a legend in the industry.  A single game would be passed through thirty-five or forty hands in a three-phase process before it was considered fit for release.  Nor were Infocom’s games neglected following their release, as is demonstrated by this note that was included in most of them: 


Here at Infocom, we take great pride in the quality of our stories.  Even after they’re “out the door,” we’re constantly improving, honing, and perfecting.  Your input is important.  No matter how much testing we do, it seems that there are some “bugs” that never crawl into view until thousands of you begin doing all those wild and crazy things in the story.  If you find a bug, or if you have a suggestion for some additional sentence syntaxes or vocabulary words, or if you found a particular puzzle too hard or too easy, or if you’d just like to tell us what you thought of the story, drop us a note! 


In response to such feedback, new versions of the company’s older games were regularly released.  Zork I went through a rather staggering seventy-some revisions over its lifetime as an Infocom product.  This devotion to bug-swatting is another Infocom principal that has been embraced by the modern IF community.  Probably the most frequently offered advice to a new or prospective IF author today is to “Beta test!  Beta test!  Beta test!” and there is even a website devoted to bringing authors together with testers willing to volunteer their time. 


            Of course, none of its technical strengths would have mattered terribly if Infocom had not employed a team of implementers, or “imps” as they called themselves, with the writing abilities and artistic vision to put the technology to good use.  Surprisingly for a company formed by a group of MIT hackers whose backgrounds were in engineering and the sciences rather than the humanities, the literary quality of the average Infocom game was shockingly good.  Further, Infocom was constantly experimenting with the forms of its games, trying to see just where it might take the seemingly limited genre of the all-text adventure game.  The level of commercial and artistic success of these experiments varied greatly, and all too often not in tandem with one another, but this restless probing at the margins is another gift of Infocom to the modern IF community.  Infocom was very much aware of the artistic potential of this new medium, and its authors felt themselves to be paving the way for others to follow.  Implementer Mike Berlyn: 


In one sense we are working within traditional genres -- mystery, fantasy,   science fiction -- and in another we are still teaching ourselves, laying out the groundwork for what these things could be.  For the most part, we are working without pioneers.  In our own way we are like Louis L'Amour or Agatha Christie or Dashiell Hammett (Dyer). 

In a 1983 Washington Post article, employee Michael Dornbrook expressed a similar sentiment: “Dornbrook believes Infocom's games are ‘the beginning of a new art form,’ one that ‘could be a significant percentage of book reading 20 years from now'” (Suplee).  Part and parcel to this attitude was a respect by the creators for their own work, something that is surprising in its rarity, then and now, in the computer games industry.  Each game’s author was given prominent billing, and the company kept its complete back catalog in print throughout its lifetime, including those of its titles that were not terribly successful commercially, and, as discussed previously, even tweaked and improved its older works when necessary.  These attitudes are closer to those of the book than the conventional computer game publisher, and, indeed, Infocom treated its products as occupying a somewhat hazy middle ground between the two.  In keeping with company research that showed virtually all of its customers to be heavy readers of books, a good portion of its distribution was done not through conventional computer game channels but rather directly to booksellers. 

              Version 3 of the Z-Machine, its first mature iteration and the one for which most of Infocom’s games were authored, was limited to story files of no more than 128K.  This is a tiny amount of space by modern standards, considerably smaller than the average digital camera photograph, but Infocom’s compression techniques made the most of it.  Even with this ingenuity, though, there was room for no more text than a typical novella.  Infocom partially surmounted this limitation by putting much of the background and atmospheric elements of its stories not in the games themselves but rather as accompanying materials in the game boxes.  This tradition began almost accidentally, when Infocom was putting together the packaging for its first post-Zork adventure, the mystery Deadline.  David Lebling: 

Well the first packaging of Zork was just the disk and the manual, very prosaic, and the first one that had really exciting packaging was Deadline, the first murder mystery we did. We had seen some things by Dennis Wheatley, I don't know what sort of books you'd call them, but they had clues, transcripts, all kinds of fun stuff in them, and I think it was Marc Blank seeing those things that motivated him to write Deadline and so we got the idea that it would be fun to have interesting stuff in our packaging too. It was such a success, and partly for that reason as well as being a good game, that the next time we did a game we thought, well, we can put some other keen stuff in it, and so we've just made a habit of it (Garret). 

 Later Infocom games would include a huge and bizarre collection of goodies to enhance the stories they told, including a starmap; a Stellar Patrol ID card; pocket fluff; a set of enchanter trading cards; a pen; a scratch-n-sniff card; a cardboard sundial; a set of bureaucratic forms in triplicate; a rubber centipede; and documents of all descriptions.  Fans took to calling these extras “feelies.”  Of course, the fact that an Infocom story could not be truly experienced and understood without its accompaniments helped to discourage piracy of the games.  In fact, Infocom frequently included information in the packaging that was essential for completing a game for just this reason.  Because the modern IF community is dependent on electronic distribution, this is one tradition that has been to some extent neglected, although some authors have chosen to manufacture feelies that may be optionally purchased to accompany their games.  A website has even been set up for the purpose of selling them.  

               In the early days, Infocom’s creative packaging extended even to its boxes themselves.  Most notable were the first releases of Starcross and Suspended.  The former was shaped like a flying saucer; the latter had an embossed robotic mask jutting from its face.  Both packages are much sought-after by collectors today, and change hands for hundreds of dollars.  However, stores were not thrilled with trying to find room for such unorthodox shapes, and the cost of their manufacture was also rather prohibitive.  Beginning in early 1984, Infocom settled on its classic “gray box” standardized package, a little marvel of compact, efficient, attractive design.  In keeping with the company’s policy of constantly updating its back catalog, older titles were re-released in this new format. 


The back of Infocom’s The Hitchhiker’s Guide to the Galaxy box,
showing the typically eclectic collection of feelies to be found within.
Note that the actual program disk appears almost an afterthought.

               At the same time that Infocom standardized its packaging, it began to categorize each of its stories by genre and difficulty level.  Games were grouped into four genres: fantasy, science fiction, mystery, and “tales of adventure,” and five difficulty levels: junior (for children), introductory, standard, advanced, and expert.  Both methods of categorization were rather problematic.  Did consumers really need to be explicitly told what cookie-cutter genre a story occupied to decide whether to purchase it?  Surely the description of the game on its box and the company’s other marketing efforts were sufficient for that decision to be made.  Further, some of Infocom’s works were too unique and innovative to be easily pigeonholed into a single genre.  Brian Mortiaty’s mournful atomic age tragedy Trinity, for instance, was arbitrarily given the label of fantasy even though it bore little relation to what the average consumer might think of when hearing that label.  The idea of rating games by difficulty makes more sense in theory, but the execution was even more flawed.  The ratings Infocom chose often seemed to bear little relation to the actual experience of players.  For instance, one of Infocom’s most straightforward early games, Infidel, was incomprehensibly rated as an advanced level title.  I remember playing that game for the first time, wondering when it would get tricky, only to find myself done with nary a head-scratch.  The less charitable might even suggest that these ratings were unduly influenced by Infocom’s marketing goals.  The Hitchhiker’s Guide to the Galaxy, Infocom’s collaborative adaptation of Douglas Adams’ best-selling novel, was rated as a standard level game, despite being laden with twisted puzzles built around Adams’ patented brand of illogical logic.  One wonders whether Infocom’s desire to attract some of Adams’ legion of fans to IF made it choose the unintimidating standard rating.  At any rate, Infocom eventually abandoned both classification systems altogether and allowed its stories to stand on their own merits. 

               A more successful innovation was Infocom’s “Invisiclues” line of hint books, one of which was made available for separate purchase for most games.  To keep the reader from accidentally revealing the solutions to puzzles, all answers were printed in invisible ink which the player must reveal with a provided pen.  This unique idea did not originate from within Infocom, but rather from a short-lived organization known as the Zork Users Group, which Mike Dornbrook founded in 1981 after a brief spell as a tester with Infocom.  On the strength of Invisiclues and other innovations, the Zork Users Group turned a considerable profit for its founder, who was thereupon hired once again by Infocom itself to serve as head of marketing.

               Infocom’s research revealed that the vast majority of its customers were repeat buyers.  A fair number bought every game immediately upon its release.  Being an Infocom player had something of the feeling of belonging to an exclusive club, a feeling partially created, deliberately or accidentally, by the company’s efforts to reach out to its customers.  It sent to everyone who had purchased and registered a game in the recent past a free quarterly newsletter, initially called The New Zork Times and then renamed to The Status Line after a certain august but apparently humorless newspaper expressed its disapproval with rumblings of legal action.  Within, along with the expected promotional materials and announcements of upcoming games, could be found humorous pieces, brain-teasers, contests, and sneak peaks at goings-on in Infocom’s Cambridge offices.  Circulation of the newsletter peaked at a very respectable 250,000.  Another unique outreach program was called “The Marathon of the Minds.”  At a public location such as a college campus, Infocom would offer an about to be released title to several teams of fans, who would compete non-stop to see who could finish it first. 

               Infocom enjoyed several very good years in the wake of the Zork Trilogy’s initial success, and racked up some very impressive sales numbers for its time.  The three Zork games sold more than one million copies together during the company’s decade in existence, and The Hitchhiker’s Guide to the Galaxy alone surpassed 250,000 in sales.  Even the average, unlicensed game could generally be relied upon to sell 50,000 copies during the company’s heyday.  A glance at the Softsel list of the top-selling entertainment software for December 12, 1983 shows Infocom holding down the top position, along with two additional games in the top ten and seven more rounding out the top forty.  The staff of Infocom was young and a bit brash, doing something genuinely new and exciting and doing well financially at the same time, and David Lebling displays some of their swagger in a 1984 Boston Globe article:  “’This business began as a lark,’ says staff writer Dave Lebling, 'and it is looking less larky every day.  We are taking serious money in, and we are putting serious money out.  All to make games’" (Dyer). 

               Certainly one thing that comes through from speaking with the implementers today and from perusing issues of their newsletter is that Infocom was a very fun place to be.  In a recent interview, Lebling described his co-workers thus: “They were very free-spirited.  More importantly, they were doing something well and having fun doing it” (Briceno).  IF scholar Graham Nelson writes that “former hands mostly look back on the heyday as a happy, one-time thing, like a summer romance” (350).  Practical jokes and assorted office hijinx abounded, yet the team there was absolutely dedicated to putting out the best product it possibly could while advancing the state of what it saw as a genuinely new art-form.  That combination of a sense of fun and an absolute dedication to not only technical but also literary quality comes through in virtually every one of the games I will discuss in detail in the next chapter. 

Sources for Further Investigation 



Zork.  David Lebling, Marc Blank, Tim Anderson, Bruce Daniels, and others. 

The portable C source of Zork’s Dungeon variant is available at http://ifarchive.giga.or.at/if-archive/games/source/dungn27s.zip

An MS-DOS executable of the same is available at http://ifarchive.giga.or.at/if-archive/games/pc/dungn27a.zip


Zork I: The Great Underground Empire.  1980.  David Lebling and Marc Blank; Infocom. 

Released as freeware by Activision, and available at http://www.csd.uwo.ca/Infocom/Download/zork1.zip


Zork II: The Wizard of Frobozz.  1981.  David Lebling and Marc Blank; Infocom. 

Released as freeware by Activision, and available at http://www.csd.uwo.ca/Infocom/Download/zork2.zip


Zork III: The Dungeon Master.  1982.  David Lebling and Marc Blank; Infocom. 

Released as freeware by Activision, and available at http://www.csd.uwo.ca/Infocom/Download/zork3.zip


Other Works: 


Addams, Shay.  “The Wizards of Infocom.”  Computer Games, February 1984.  

Available online at http://www.csd.uwo.ca/Infocom/Articles/wizards.html


Anderson, Tim and Stu Galley.  “The History of Zork – First in a Series.”  The New Zork Times, Vol. 4, Num. 1, Winter 1985.  Available online at  



Anderson, Tim and Stu Galley.  “The History of Zork – Second in a Series.”  The New Zork Times, Vol. 4, Num. 2, Spring 1985.  Available online at  



Anderson, Tim and Stu Galley.  “The History of Zork – The Final (?) Chapter: MIT, MDL, ZIL, ZIP.”  The New Zork Times, Vol. 4, Num. 3, Summer 1985.  Available online at http://mirror.ifarchive.org/if-archive/infocom/articles/articles.zip


Blank, Marc S. and S.W. Galley.  “How to Fit a Large Program into a Small Machine.”  Creative Computing, July 1980.  

Available online at http://mirror.ifarchive.org/if-archive/infocom/articles/articles.zip


Briceno, Hector, Wesley Chao, Andrew Glenn, Stanley Hu, Ashwin Krishnamurthy, and Bruce Tsuchida.  Down From the Top of Its Game: The Story of Infocom, Inc.  

Available online at http://mirror.ifarchive.org/if-archive/infocom/info/infocom-paper.pdf


Dyer, Richard.  “Masters of the Game.”  The Boston Globe Magazine, May 6 1984.  

Available online at http://www.csd.uwo.ca/Infocom/Articles/globe84.html


Garret, Roger.  “Infocom Special.”  Your Computer, vol. 7, no. 3, May 1987.  

Available online at http://www.csd.uwo.ca/Infocom/Articles/yc87.html


Infocom Advertisement.  “We Stick our Graphics Where the Sun Don’t Shine.”  Analog #11, April/May 1983.  

Available online at  http://mirror.ifarchive.org/if-archive/infocom/adverts/ad1.jpeg


Lebling, P. David.  “Zork and the Future of Computerized Fantasy Simulations.”  Byte, December 1980.  

Available online at  http://mirror.ifarchive.org/if-archive/infocom/articles/articles.zip


Lebling, P. David, Marc S. Blank, and Timothy A. Anderson.  “Zork: A Computerized Fantasy Simulation Game.”  IEEE Computer, vol.12, num. 4, April 1979.  

Available online at http://mirror.ifarchive.org/if-archive/infocom/articles/articles.zip


Nelson, Graham.  The Inform Designer’s Manual.  4th ed.  The Interactive Fiction Libray:  St. Charles, Illinois, 2001.  

Available online at http://ifarchive.flavorplex.com/if-archive/infocom/compilers/inform6/manuals/designers_manual_4.pdf


O’Brian, Paul.  “Archeology.”  IF-Review, June 29 2001. 



Sherwin, Robb.  “nevermind.”  Usenet post to rec.games.int-fiction, June 5 2001. 


Suplee, Curt.  “Through the Zorking Glass.”  Washington Post, December 22 1983.  

Available online at http://www.csd.uwo.ca/Infocom/Articles/post83.html. 


<<- Chapter 3   Table of Contents   Chapter 5 ->>