By 1981 Adventure was widely known among programmers and, as described in the next chapter, it was also widely imitated by them. Tracy Kidder's Pulitzer Prize-winning Soul of a New Machine brought it to the attention of the general reader that year. The book described, probably better than any academic article or piece of journalism since then, how a particular interactor actually approached a work of interactive fiction:
Carl Asling's cluttered little area made a small rectangle of light. Strewn before me across the surface of his desk, like relics from a party, lay dozens of roughly drawn maps. They consisted of circles, inside of which were scrawled names such as Dirty Passage, Hall of Mists, Hall of the Mountain King ... Webs of lines connected the circles, and each line was labeled, some with points of the compass, some with the words up and down. Here and there on the maps were notations-"water here," "oil here," and "damn that pirate!" (Kidder 1981, 86)
Making such maps was an essential part of solving Adventure and would remain essential to interacting with most other works of interactive fiction. Kidder also described how, during an after-hours introduction to Adventure, more than one person was involved in interacting. In this case, Kidder typed while Asling looked at what he was doing and commented. This was one observation of people, playing this cooperative game, who did not interact in solitude.
Adventure has been the topic of a good deal of discussion and analysis, even if many important questions about it remain. This short section has treated only the most important aspects of the history and creation of Adventure. Despite the uncertainties that remain about exact dates and about the differences between the original and expanded works, a great deal is known about the influences on Adventure, the contexts in which it was created, and the authors themselves. While Crowther's original program is not available as of this writing, the canonical Adventure as expanded by Woods is one of the most universally available computer programs in existence, running on hundreds of platforms.
As important as Adventure is, however, it is impossible to learn about the course of interactive fiction's development by looking in detail at only this first work; the development of the form must be seen in how later works differ from Adventure. The next chapter undertakes a detailed comparison of Zork and Adventure, with a focus on the former work. Aarseth (1997) complains, with justification, that Zork "is often, and undeservingly, claimed as the paradigmatic adventure game" (108). (For instance, Hamlet on the Holodeck, the first scholarly book to discuss Zork at length, does not mention Adventure U. Murray 1997).) In this book Zork is treated in more detail than Adventure to illuminate the form's history, not to set up Zork as the paradigm for the form. Zork was certainly important and popular, however; it and the other mainframe works that followed would serve to show that Adventure was not a dead end.
For those who had access to mainframe computer time, writing one's own Adventure-like program seems to have been almost as popular an activity as was playing Adventure. Thus, Adventure became the model for computing gaming overall; Nelson (2001b) writes that "for the five years to 1982 almost every game created was another `Advent"' (347). At Stanford, where Don Woods had augmented Will Crowther's original work to create the canonical Adventure, at least two follow-ups were written on different systems during this time: FisK and Lugi (Meier 2002). The authors of these two works are not recorded online, and the source code has not been made available. At the University of Waterloo in Canada, Marc Niemiec created a language called F (for Fantasy) for the development of interactive fiction on the Honeywell Level 66, using it to write New Adventure. Brad Templeton and Kieran Carroll wrote Martian Adventure in F, incorporating at least one puzzle that was somewhat "Martian" in the sense of the Martian Poetry movement, and shows the relationship between the riddle and early interactive fiction: the player character encounters "keys" that cannot be picked up. It turns out these are not the sort of keys one expects in an adventure; to deal with them one is supposed to type (Meier and Persson 2002). For the Hewlett Packard 1000, two IF works were programmed and made freely available: Mystery Mansion by Bill Wolpert and Warp. An incomplete adaptation of The Lord of the Rings entitled Lord, created at the University of Helsinki by Olli J. Paavola, has been called, by Nelson (2001b), "the first book adaptation in interactive fiction" (347). Nelson noted that Lord, influenced by Zork as well as Adventure, was particularly authentic in the way it created a world with the sort of detail Tolkien would appreciate.
At Carnegie Mellon University, during 1979-1982, John Laird (1997) wrote the Adventure follow-up Haunt, which he has described as "over-thetop," "quirky" and also "a bit buggy." Nelson calls it "not inspired" but notes that it is the first non-cave work of interactive fiction (2001b, 353).The outrageous prologue is worth quoting from extensively. It merely introduced the sort of treasure-hunt situation that quickly became stereotypical, but it is a great departure from Adventure and Zork, which simply begin with the player character in the woods.
Along [sic] time ago, a young couple was picnicing [sic] near the woods on the outskirts of town. They were celebrating the birth of their first child. Unfortunately, a crazed moose inhabited that area and attacked them. The child and husband were unharmed, but the wife was gored to death by the moose.
After the funeral, the man bought the land where the incident occurred and constructed a large mansion: CHEZ MOOSE. He filled it with the treasures of his family and claimed that his wife's soul was still in the area. He vowed to remain in the mansion until he had returned her soul to human flesh. He tried to bridge the gap between life and death to reclaim her. . . . Several people have entered the mansion looking for him but none of them have ever returned. There were rumors that he and his wife now haunt the house.
That would be the end of the story except that the house still stands and is filled with priceless treasures. The house and all its contents are willed to his only descendant.... The terms of the Will say you get to keep any treasure you get to the lawn, but of course you must also get off the premises alive.... If you are insane enough to try, your adventure starts at a bus stop.
Laird (1997) himself wrote that "the puzzles were way too obscure (many based on Saturday morning cartoons from my youth)" but admitted they were certainly creative, suggesting a different, even wackier mode for interactive fiction.
One puzzle involves removing the paint from a painting to reveal a more valuable one underneath. Another requires the player character to use two straight objects to make a vampire-resisting cross.
Numerous obscene inputs are recognized by Haunt. Marijuana is one of the treasures; smoking it makes the adventurer hungry. Stranger effects are brought on by eating the sugar cube. After winning the house, the player character is confronted with a final dilemma when James Watt arrives from the Department of the Interior and asks to buy the land. Written in CMU's OPS-4 (Official Production System 4), Haunt made its way to a few other universities but was not widely available. It was notable for being the first rule-based system to have more than 1,000 rules (Laird 1997). Laird (1998) wrote it without having played Zork; the creators of Zork played it at some point before Zork was finished (Nelson 2001,353; Lebling 1997). Laird began a port of Haunt to OPS-5; this incomplete version has been used as an exercise by CMU students (Winalski 1997). CMU would later host the major academic effort in interactive fiction, Joseph Bates's Oz Project, discussed in chapter 7. Laird, now a professor of computer science at the University of Michigan, is currently developing a new domain for his artificial intelligence research, which he describes as "an adventure game where the player takes on the persona of a ghost-like energy creature trapped in a house" (Laird 2001).The graphical, 3-D system is being built using the Unreal Tournament engine and is known as Haunt 2.
The mainframe Zork, programmed at MIT beginning in May 1977, has a special place among the mainframe follow-ups to Adventure. Zork became the second widely known interactive fiction work. For less capacious home computers, it was split up and further developed into Zork I, Zork II,
and Zork III -a trilogy of best-selling games that was the basis for Infocom's creative and financial success, which I discuss in chapter 5. Several specific improvements that had been made to the Adventure world model and parser were touted in an academic paper (Lebling, Blank, and Anderson 1979), and, later, in the manuals that accompanied Infocom's commercial products. Those familiar with the history of computer gaming recognize the significance of Zork. One contemporary IF author expresses a common view: "It's by far the most famous piece of IF and can be considered the father figure of the genre, much like Super Mario Bros. is with the side scroller or King's Quest for the graphic adventure" (Sherwin 1999). One rough measure of the overall popularity of Zork is that in mid-December 2001, the word "Zork" appeared in approximately 124,000 Web pages in the Google index; only about 22,000 such pages mentioned either "Space War" or "Spacewar," referring to the MIT program that is widely considered to be the first modern video game. Even taking into consideration that the word "zork" sometimes appears on the Web in its original sense-a nonsense word in MIT argot-the interactive fiction Zork is clearly well known; it is also important in the history of new media and of interactive fiction.
Zork in this chapter refers to the PDP-10 computer program, running on ITS and written in MDL, that was developed at MIT beginning in late May 1977 and augmented later that year and in the following years (Anderson 1985a, 7). For a few months Zork was called Dungeon; there is a version of a FORTRAN port of Zork that has been widely available since 1979 that uses this name. Zork also refers to ports from that FORTRAN port into Glk and from the original MDL into Inform. Zork is distinguished in this discussion from Zork I-III. I consider that trilogy and how it differed from the mainframe Zork in the next chapter. Where distinctions among different versions of Zork are important, the year of the version cited is given.
Despite the clear importance of Zork, with very few exceptions, the advances that this work brought about have been only superficially considered. This chapter describes three advances and their importance. First, Zork draws on a deep reservoir of technical humor, making its origins in the subculture of MIT evident and commenting on technology in interesting ways. It first realized the ability of interactive fiction to speak back to the culture in which it was produced. Second, while the technical advances made by the game's creators (known as Implementors) were in some ways incremental, they did lead to progress in interactive fiction. It is helpful to explain these advances and see how these proved important in the development of later works. Third, Zork was at least in some ways superior to Adventure as a system that generates satisfying narratives in response to user input. It produced narratives that were connected to adventure stories of the folktale sort in new ways-not just through offhand reference, but structurally. Neither programming improvements nor better writing, considered alone, made Zork's thief the first memorable character in interactive fiction; rather, it was the way he functioned during an interaction and throughout a successful traversal of Zork that was fundamentally better than what had been achieved before.
SEEING ZORK AGAINST ADVENTURE
The potential narratives of a successful traversal of Zork involve the adventurer wandering through an outdoor area, a house, and then a vast complex called the Great Underground Empire.To succeed the interactor must direct the adventurer to collect treasures, almost all of which are located underground. A handful of living opponents thwart the adventurer: the troll, who stays put in a single room and serves as an obstacle; the vampire bat, who can carry off the adventurer; the cyclops, who can dine on the adventurer; and the thief, who wanders around the underground areas stealing items the adventurer either is holding or has already seen. To get through the mazes, detailed mapmaking (or else extraordinary luck, or cheating) is required. Riddle-like challenges require the interactor to understand the nature of a disguised object in order to use it properly; in some cases machinery must be manipulated in order to determine its purpose. Some puzzles, such as the Bank of Zork and the Royal Puzzle, are elaborate and extremely hard. A few require that the interactor "guess the verb" and perform an action that would not be obvious from the commands available in Adventure. If a magic word from Adventure is typed in Zork, a hollow voice says "Cretin."
It makes little sense to consider certain features in Zork as if they were original, because certain features are not. Attacking Zork with a detailed psychoanalytical reading that considers the subterranean world, beneath a forested area in which a lone building sits, is foolhardy-for several reasons, but most specifically because all of these features are lifted directly from Adventure. According to the small leaflet found in the mailbox in the 1978-1979 FORTRAN port of Zork, the program "was created ... by Tim Anderson, Marc Blank, Bruce Daniels, and Dave Lebling. It was inspired by the Adventure game of Crowther and Woods, and the Dungeons and Dragons game of Gygax and Arneson." By 1981 the leaflet had been revised to replace mention of Dungeons and Dragons with "the long tradition of fantasy and science fiction adventure." This revision came after Zork was briefly renamed Dungeon, drawing the attention of certain lawyers to the Implementors and MIT (Anderson 1985b, 4). Both the original and revised statements have truth to them, of course.
Zork's Implementors had played Adventure earnestly, and they created Zork mainly in response to their enjoyment of that interactive fiction work and the opportunities they saw to improve upon it. The interests of the Implementors were not restricted to programming. As Anderson (2001) said, "One reason that some of the writing is good, if it is good, is that the people who wrote it were not nerds in the classic sense. Very few people are really nerds in the classic sense, I think. Marc, Dave, and I all exposed ourselves to a lot of really good writing."The Zork Implementors were also filmgoers, seeing Star Wars very soon after it came out, during Zork's early development; Marc Blank was on MIT's Lecture Series Committee, the student group that screens films on campus, and had an interest in classic films. The Implementors read books and viewed films eclectically, taking in much from outside the genres of science fiction and fantasy. This included history, as one might guess from the offhand reference in Zork to Octavian's general Marcus Vipsanius Agrippa.
Few people are struck by the powerful narratives that Zork produces when compared to those produced by Adventure. Zork, like Adventure, provided nothing special to set up the initial situation, in which the adventurer is standing near a building in the woods; certainly Haunt was much more innovative when this particular aspect of writing is considered. Although the humor of Zork is widely appreciated, in some ways Zork's writing may in fact be worse than the writing in Adventure, especially in the case of the accuracy and richness of geological description. Nelson (1995a) described Zork's underground setting as "based not on real caves but on Crowther's descriptions" and as "better laid out as a game but not as convincing." Lebling (1988) explains that those at MIT who weren't cavers still had their own underground empire to explore and to base interactive fiction upon:
When I was a student at MIT, there was a pastime called "Institute Exploring" (also known as "Tunnel Tours"). A group of students would go over to the main part of the campus at around 3am and try to visit some of the more obscure and off-limits locations.
MIT is full of basements and subbasements, and these are often crammed with equipment left over from some cancelled research project.
However many levels of simulation are involved in the room descriptions, though, Zork did substantially improve upon the art of interactive fiction. Janet Murray (1997) describes one way in which Zork presents a compelling "moment of experimental drama," early on: when the adventurer first enters the dungeon, the trapdoor is closed and the way out is barred by an unseen opponent (81-82). This moment, which has no parallel at the beginning of Adventure, creates an ominous and interesting situation in the potential narratives of Zork. It also makes for a puzzle-the interactor must figure out how to escape, going deeper into the dungeon.This improvement, one of several that made interactive fiction and new medi
a more compelling, came about by improving the design of the IF world, not by simply programming more cleverly or by writing better bits of texts.
Zork is a literary, gaming, and computing artifact, a part of the culture and subculture in which it was created. It has been noted that masculinist rationality underlies Zork and other interactive fiction (Sloane 2000, 100); other contexts of interest also exist. The Implementors were certainly influenced by, but not mere puppets of, male academic research and its perspectives and concerns. In Zork they sought to comment on academia and technology as technologists and as writers and designers of a new kind.They were marked in certain ways by their environment, and they marked back in Zork, the simulated environment they created.
To illustrate how an MIT-specific interpretation can help us understand Zork, consider that in the Entrance to Hades a large gateway bears this inscription: "Abandon every hope, all ye who enter here." Relating this inscription to literary tradition is, if not fruitless, misleading. Zork's Hades has only superficial similarities to either Dante's Hell or the Hades of the ancient Greeks, with jeering "evil spirits" who prevent entry. Dante's Hell lies conspicuously open and unguarded-in contrast to the securely locked gate of Purgatory, watched over by a sword-wielding angel (Purgatorio 9.76-138). Christian doctrine of Dante's time held that it was easy to enter hell; entering Zork's equivalent realm is difficult, providing an intricate puzzle opportunity. A different approach to this Hades entryway is provided by Implementor Tim Anderson (2001): "You can think of it as a Dante quotation or you can think of it as something people say when they come to MIT, which is really more what we had in mind." MIT students come to learn that "Tech Is Hell," and admission to that Hell is indeed difficult. There was also a rather direct antecedent in computing, whether or not it was known to the Zork Implementors: The (rather arcane) source code for the Unisys A Series COBOL 68 compiler began with a comment containing the same quotation, in Italian (Stevens 1999).
Twisty Little Passages Page 12