The Online Interactive Fiction Review Site
Dungeon is the predecessor to the Zork games; it was MIT's answer to Crowther and Woods' "Adventure" and, much like that game, it lived on a mainframe, since its prodigious size was too great for the personal computers of its day. When the authors decided to make a commercial go of the text adventure business, they chopped Dungeon into three sections, rearranging the geography and adding some new elements to each chapter, especially the second and third. I've played the Zork games many times, but I had always wanted to play the mainframe version in order to better understand just what was added and what subtracted. So when I opened the WinGlk version of Andrew Plotkin's C translation of the game, I was prepared for some shifts in layout compared to my deeply-graven memories of Zork I.
What I wasn't prepared for, though, was the way in which Dungeon gleefully confounds any sense of actual geography in exchange for making the game map another obstacle to be overcome. In Dungeon, connections that line up properly (for example, leaving one room to the south and entering the adjoining room from the north) are the exception rather than the rule. Instead, you may go west and find that to get back to where you came from, you have to go west again. In a recent article about crafting a good setting for fantasy IF , Emily Short addressed this tendency:
[In] the ideal IF setting, the parts of the setting relate to each other in comprehensible ways. Things are located sensibly. I dislike mazes not only because you do have to map them but also because they interfere with and scramble up the intuitive sense of place that I otherwise build up as I play.
In this sense, almost the entirety of Dungeon functions as a maze, and any coherent sense of place that might emerge is bound to get smacked down as soon as the next exit is explored. I have a pretty good knack for mapping in my head, and thus don't tend to make a map while playing IF but, with this game, there was no way I could pursue that strategy. Thus, grumbling, I hauled out my copy of GUEmap and tried diligently to record the tortured web of interconnections that make up the Dungeon landscape. When I finally finished, I uploaded the results to GMD so that other players like myself won't have to struggle through the game's mazes on their own.
And oh, the mazes -- in addition to the general illogic of its structure, Dungeon also sports several mazes, all of which carry the "warped connections" tendency to its furthest extreme. Of course, when seen from the historical perspective, these mazes make sense: Adventure had mazes, and since mazes are one of the easiest kinds of puzzles to create, it follows that the game attempting to top Adventure would have mazes of its own. What Dungeon does, though, is to twist the knife: not only does it present the player with mazes, it confounds the typical "drop item and map" strategy by having an NPC come along and remove or rearrange those items, taunting the player with comments like "My, I wonder who left this fine hot pepper sandwich here?"
When viewed with a modern eye, obstacles like this make clear how different is the stance of modern IF from its ancestors. 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. As Robb Sherwin asserted on rec.games.int-fiction recently , "Zork hates its player."
Today's IF, by contrast, works a bit harder to collaborate with the player, with the aim of creating a shared experience, both in setting and plot. Even the Zork games moved in this direction, at least in comparison to Dungeon, mitigating some of the latter game's greatest excesses by straightening out many map connections, allowing more flexibility with the permanent light source, and providing a bit more description from time to time. The ways in which Infocom itself engineered the shift from "text-based puzzle games" to actual interactive fiction is a subject for another article, but what's become clear is that where the emphasis was once on opposition, it has shifted steadily to cooperation.
To my mind, this shift is both appropriate and necessary, and what playing Dungeon illuminated for me is that this movement towards collaborative IF is not the same thing as the concurrent movement towards "literary IF", though they are often confused for one another. I can envision a game that, like Dungeon, has no particular literary pretensions, but unlike Dungeon, isn't trying to undermine its player through the use of arbitrary techniques like twisty map connections and unreliable light sources. I would assert that collaborative IF doesn't need to tell a story, and it certainly doesn't need to aspire to literary greatness, but it does need to work with the player to create a rich, interactive world, and it does need to be concerned with giving the player a positive, fun experience. Of course collaborative IF can be puzzleless, but it needn't be -- puzzles can be part of the fun, as long as they aren't geared towards forcing restarts after 800 moves, or making the player do tedious, menial work. The move away from antagonistic IF is the reason why things like mazes, limited light sources, and starvation puzzles are met with a chorus of jeers these days, but the elimination of these elements doesn't necessarily dictate anything in particular about how literary or puzzleless a game might be. Instead, the change makes the whole experience of IF more about fun than bloody-minded perseverance; playing Dungeon makes it clear how necessary this change was, and how far we've come since those mainframe days.
 Short, Emily. "Developing A Setting For Fantastical Interactive Fiction", 2001.
 Sherwin, Robb. "Re: nevermind". rec.games.int-fiction, 2001/06/05