Infocom helped establish many conventions of interactive fiction. As is often the case in art and literature, some of the most interesting work in recent years has sought to subvert these conventions and use the text adventure interface for novel purposes. As one technology writer explained in a subtly mangled metaphor that rings oddly true, technologically and artistically, "Infocom ... broke important new ground that is still being explored today" (Garfinkel 1999, 83). Indeed, Infocom began work on the foundation of interactive fiction while the plot of ground that it was to be built upon had not yet been completely surveyed.
The thirty-five text-based Infocom games (see table 5.1) were all published during the 1980s. The early works, which contained no graphics at all, were available on a tremendous range of platforms. Infocom's publications spanned several popular literature genres, moving beyond the fantasy category into detective fiction, science fiction, modern-day treasure-plundering excursions, children's fiction, the spy thriller, and historical fiction/romance. A few of these IF works seemed to have more high-cultural aspirations (which were achieved in part) but almost all could be reasonably fitted into some popular genre, and Infocom did just that by conspicuously labeling them in the company's later years. Near the end of Infocom's independent existence, they created a few works based on classic stories, as many of their competitors had done.
Infocom's works, central as they were to the development of interactive fiction, were not known worldwide in the 1980s. For instance, Nelson reports that "Infocom's wares were rarely-seen exotica in England: a luxury brand for those in the know" (2002) and notes that while "Infocom was dominant for a period in the higher-end, chiefly American market ... the company was not nearly so visible outside the USA, where disc drives were less affordable" (2001b, 350). Infocom's influence is indeed seen today in the worldwide IF community, but it did not extend across even the Englishspeaking world in the 1980s. Other important companies that were contemporaries of Infocom, inside and outside the United States, are discussed in the next chapter.
Although Infocom was early to market and became the most important company of the commercial era, it was not the first to sell microcomputer text adventures. That honor goes to Scott Adams-no relation to the creator of the Dilbert comic-who released a tiny BASIC version of Adventure in 1978. He wrote the program, Adventureland, on his TRS-80 Model I. After selling a few copies of the cassette tape, he published it in a packaged format through the company he founded, Adventure International. His next work was Pirate Adventure, which was more original and cohesive. Its source code was published in the December 1980 issue of Byte, making Adams's work, and interactive fiction in general, more visible. His company would publish thirteen other text adventures before going bankrupt in 1985 (Granade 1999a).Written for cartridge- and cassette-based computers with very little memory, his works have been described as "strong on imagination and filled with clever puzzles. Their main weakness is rather poor English usage and spelling" (McGath 1984, 45).These works are not known for their literary qualities. As Nelson (2001b) reports, their descriptions had "weirdly errant grammar" (348) and seemed to be written "in pidgin English, like telegraphese" (200).
TABLE 5.1
The thirty-five canonical interactive fiction works developed at and published by Infocoin
Arthur: The Quest For Excalibur by Bob Bates (1989)
Young Arthur must find the pilfered sword-in-stone. Graphical; Infocom's last. Ballyhoo by Jeff O'Neill (1986)
A circus visitor snoops to discover a kidnapping mystery. A darker, wry humor. Beyond Zork: The Coconut of Quendor by Brian Moriarty (1987)
Had hit points (like Dungeons and Dragons), an on-screen map, and a rich world. Border Zone by Marc Blank (1987)
A spy thriller in a made-up East European country. Events happen in real time. Bureaucracy by Douglas Adams (1987)
Absurd red tape and procedural challenges are joined with Adams's wit. Cutthroats by Michael Berlyn and Jerry Wolper (1984)
A diver recovers an undersea treasure with help from unreliable team members. Deadline by Marc Blank (1982)
A detective interviews suspects at a family house to find the murderer. Enchanter by Marc Blank and Dave Lebling (1983)
A novice spell-caster must defeat a warlock. A follow-up to the Zork Trilogy. Hitchhiker's Guide To The Galaxy by Douglas Adams and Steve Meretzky (1984) Arthur Dent explores space in his bathrobe.A clever reworking of Adams's book. Hollywood H~jinx by Dave Anderson and Liz Cyr-Jones (1986)
The will calls for a treasure hunt at the dead Relative's California mansion. Infidel by Michael Berlyn and Patricia Fogleman (1983)
Abandoned by the dig team in the desert, an archaeologist plunders a pyramid. James Clavell's Shogun by Dave Lebling (1989)
In 1600, an Englishman enters a feudal Japan straight from the book. Graphical. Journey: The Quest Begins by Marc Blank (1989)
A party of five go adventuring. Graphical and with role-playing game elements. Leather Goddesses of Phobos by Steve Meretzky (1986)
A space opera with an impending alien attack and more bondage than most dungeons. The Lurking Horror by Dave Lebling (1987)
During late-night studying, alchemical secrets are uncovered at an MIT-like campus.
A Mind Forever Voyaging by Steve Meretzky (1985)
Several dystopian trips to explore the possible futures of a city. Moonmist by Stu Galley and Jim Lawrence (1986)
A night in a haunted castle gives a visitor the opportunity to uncover mysteries. Nord and Bert Couldn't Make Head Or Tail Of It by Jeff O'Neill (1987)
Wordplay replaces the usual adventure actions in saving the town of Punster. Planetfall by Steve Meretzky (1983)
The robot Floyd helps in a comic sci-fi quest for rescue from a strange planet. Plundered Hearts by Amy Briggs (1987)
A romance-novel adventure with pirates, swords, and a female player character. Seastalker by Stu Galley and Jim Lawrence (1984)
A monster threatens an undersea research lab. The first for younger interactors. Sherlock: The Riddle Of The Crown Jewels by Bob Bates (1987)
Watson, accompanied by Holmes, takes the lead in recovering the crown jewels. Sorcerer by Steve Meretzky (1984)
An evil demon menaces a land whose most powerful enchanter is missing. Spellbreaker by Dave Lebling (1985)
The head enchanter faces the final crisis: the breakdown of magic itself. Starcross by Dave Lebling (1982)
An alien craft holds logical challenges. Hard, and in the genre of hard sci-fi. Stationfall by Steve Meretzky (1987)
A Planetfall sequel aboard a space station that's deserted-except for Floyd. Suspect by Dave Lebling (1984)
The suspect must prove that someone else committed the murder at the masked ball. Suspended by Michael Berlyn (1983)
A quiescently frozen controller directs robots to save a far-future complex. Trinity by Brian Moriarty (1986)
A fantasy world offers passage back in time and a way to stop nuclear catastrophe. Wishbringer by Brian Moriarty (1985)
A daydream turns real and a letter carrier quests through a magically changed town. Witness by Stu Galley (1983)
Blackmail turns to murder in this hard-boiled detective game, set in the 1930s.
Zork I:The Great Underground Empire by Marc Blank and Dave Lebling (1980)
The generic adventurer visits a dazzling, below-ground kingdom-and loots it. Zork II: The Wizard of Frobozz by Marc Blank and Dave Lebling (1981)
Another segment of the underground kingdom holds an annoying, spell-casting wizard. Zork III: The Dungeon Master by Marc Blank and Dave Lebling (1982)
The final dungeon segment is intricate, with perhaps the hardest Infocom puzzle. Zork Zero: The Revenge of Megaboz by Steve Meretzky (1988)
A silly object-collecting romp through a pre-Zork underground. Graphics.
Language understanding was limited, too, but the terse descriptions and two-word parser allowed this limited sort of interactive fiction to fit into 16k of memory; frequently repe
ated phrases, which only had to be stored in memory once, also helped to trim down the size of programs. The player character was conceived of somewhat differently, as a reviewer noted: "the Adams programs let you give commands to a more or less obedient `puppet.' Instead of saying `you are at the top of a tree' ... the program says `I am at the top of a tree"' (McGath 1984). Adams (2002) described his software development process and how feedback from testers was important to it:
I'd set my theme, I'd set my locations, and I'd start putting items in and putting in puzzles. I'd get the game about two thirds done and then I would stop. The next one third of the game literally came from the people I gave to play the game. I'd watch how they played the game, I'd watch what they tried to do with the items that I never thought they might try to do.
Although not very interesting to consider either as an interactive system or as potential literary art, Adams's works were widely available and influential. They remain good examples of how to cope with very limited computer resources. Many interactive fiction authors active today (e.g.,Adam Cadre) first learned of the form through Scott Adams's works. The term interactive fiction itself was first popularized by Adventure International. It seems to have been coined by Robert Lafore, who started a company called Interactive Fiction and produced BASIC programs for the TRS-80, including Six Micro Stories, which were distributed by Adventure International (Lafore 2002). Although Six Micro Stories generated some critical enthusiasm for the possibilities of the form (Niesz and Holland 1984), it was exceedingly primitive, with long passages of text punctuated occasionally by a binary choice. The user could type in an English reply in order to indicate this choice, but formally, as potential literature, these programs had more affinity with Queneau's "Un Conte a votre facon" or with a Choose Your Own Adventure book than with Adventure and the interactive fiction works that followed it, those works that simulated complex worlds beneath the generated texts.
The type of entertainment software development that Adams and Adventure International exemplified-the single person serving as programmer and designer who worked with software testers-would change dramatically during the 1980s, as project budgets swelled and the development of games became a more complex affair. With some notable exceptions, commercial interactive fiction would, through the 1980s, follow the Scott Adams model, with a single person serving as programmer and writer. This was even the case at the flagship company of the form, founded by a group of programmers not long after Adventure International was started.
INFOCOM'S BEGINNINGS
Infocom was founded on 22 June 1979 by alumni and other affiliates of MIT's Project MAC (Multi Access Computing/Machine Aided Cognition), the group that later became the Laboratory for Computer Science. Three of the four Zork creators (Marc Blank, Tim Anderson, and Dave Lebling) were founding members of Infocom. There were seven other founders, including Joel Berez, Mike Broos, Scott Cutler, Stu Galley, and Chris Reeve.The other two were AlVezza, head of the Dynamic Modeling group at MIT, and J. C. R. Licklider,Vezza's mentor and the founder of Project MAC. Infocom was not chartered as an interactive fiction company. The founders put the company together to sell some type of money-making software and planned to figure out exactly what that would be later. In August 1979, Blank became director; that November, Berez was elected Infocom's first president.
Berez and Blank became the first paid employees in June 1981; they were paid, at first, in IOUs. But as the early Infocom games began selling well and the company obliged customers with more, Infocom grew rapidly: by mid-1985 sales exceeded $1() million and the number of employees had topped one hundred. Before the close of the decade, however, fortunes had reversed and a tiny remnant of Infocom-five people-was all that was left. Those five moved to California to work at the headquarters of Infocom's new parent company. Despite the extremes that Infocom went thorough, when compared to the heady successes and abrupt downfalls of recent Internet start-ups or earlier video game makers such as Atari, this corporate rags-to-riches-to-rags story is not particularly dramatic. Considered in the context of interactive fiction, however-and as part of the overall intellectual history of computing-the rise and fall of Infocom is certainly of importance and interest.
The widespread availability of Infocom games was important to Infocom's commercial success during the era of the home computer, when a huge array of platforms were in use and no single one dominated the market. Cross-platform availability also broadened the audience for text adventures to include the users of business computers and other systems that were traditionally not seen as gaming platforms. Infocom didn't code up different versions of each work for Apple and Atari and Commodore, as many game makers of the 1980s did. Implementors were able to develop a single program to be used on any platform; they worked on terminals connected to the company's DEC System 20 using a special development system (Addams 1984). In order to deploy their games across different platforms, Berez and Blank specified and programmed a virtual machine called the Z-machine. This software computer could be implemented on many different platforms, including almost all of the popular microcomputers in the United States during the 1980s. (The only exceptions were very early or severely underpowered machines, such as the Commodore PET.) This approach was not only efficient with memory, it was also what allowed Infocom to develop for only one platform: the Z-machine. Since the Z-machine, actually software itself, was then widely implemented on many home computers (and could be implemented on new computers as the need arose), any new interactive fiction creation could instantly work on every major platform and many minor platforms. Developing twenty implementations of the Z-machine and twenty IF works would result in four hundred saleable products. Mike Dornbrook (2000), Infocom's director of marketing, said this cross-platform availability was important to the company's bottom line: no single computer platform ever accounted for more than 25 percent of Infocom's revenue in any quarter.
The Z-machine served as the interpreter for Infocom story files. To create the story files themselves-the code that the Z-machine ran-a compiler was programmed by Blank to convert high-level, more English-like instructions into executable form. The high-level language that was used was an Infocom original, based on the Lisp-like language MDL, which had been used to write the mainframe Zork. The new language was called ZIL, Zork Implementation Language. Here is the ZIL for the living room in Zork I:
From the standpoint of IF developers, ZIL made several advances over MDL (Galley 1985). For one thing, it replaced what one developer called "moreor-less incomprehensible verb definitions" with more legible ones (Lebling 1996). The virtual machine concept was first implemented commercially in the Z-machine. (The earlier, noncommercial implementation was UCSD's P-Machine, which ran Pascal (Foust 2001).) This same concept is what enables Java programs to work across different platforms by running in Java virtual machines.
Infocom's successful product line was based on the mainframe Zork, a group project. Unsurprisingly, Infocom became a collaborative environment. Although one Infocom author said "We don't brainstorm as much as you might think" (Moriarty 1987), the author or authors listed on the packaging were seldom the only ones who helped to shape the games. One journalist who visited the company wrote that "a single writer originates each project, but the process soon becomes collaborative, with solutions and scenarios proffered, swapped, or stolen" (Suplee 1983). Also, a new IF work would be developed by starting with the code from an earlier project and modifying it; according to one Infocom IF creator, "5 to 10% of a game's total code comes from the `generic' shell" (Moriarty 1987). This reuse, although a small part of the overall program, helps to explain a feature of IF works noted by Nelson (200 1b)-namely, that "the strongest unity of style between the Infocom games is that they seem to be told by essentially the same narrator" (354). Usually either one individual or a team of two, working in slightly different roles, would head up the creation of a particular interactive fiction piece. One person would often serve in what would
, in some other organizations, be several roles: as the main writer of the texts involved, the designer of interactivity and puzzles, and the ZIL programmer. "Too little credit has been given to department heads who were at least as responsible for Infocom's artistic texture" (350), Nelson noted. These included Jon Palace, a former book editor who managed the Implementors, and the heads of quality assurance: first Dave Anderson (who then became an Implementor), then Liz Cyr Jones (Lebling 2002).
Infocom made its first sale in 1980: an 8-inch floppy disk containing the PDP-11 version of Zork I. To make the sprawling Zork fit on personal computers of the era, Blank and Lebling had implemented only one somewhat embellished segment of the game in their first commercial program. (At that time, home computers usually featured about 32 KB of RAM and one 100 KB floppy drive. The PDP-11 was a more powerful computer, but Zork I was developed on the Z-machine, which would work on less powerful platforms as well.) Blank and Lebling finished Zork II in 1981 and Zork III the following year, incorporating elements from the mainframe work but also adding more original text and code.The modifications make these three games new works, distinct from the mainframe Zork and different from one another. Infocom initially sought distribution for its first products through other companies-one implementor said that "by the time Zork fan Bill Gates heard of our offer, Infocom was deep in negotiations with Personal Software Inc." (Galley 1985)-but it chose to publish the works itself after a short period in which Zork I and Zork II were distributed through Personal Software.
The fame of Zork I eclipses its actual importance to interactive fiction. That game brought the more advanced elements in mainframe Zork to the home computer user, to be sure, and it became the basis of Infocom's commercial success.Yet Infocom's first release, when compared to its mainframe ancestor, features only slight improvements, whether considered in traditional terms of puzzlecraft and literary merit or seen in the context of world and riddle.
Twisty Little Passages Page 15