My good friend Harry Kaplan recently sent me a link to a little Oregon Trail short story by Caitlin Horrocks. The whole joke here — and it’s a cute one — is in the fun it pokes at the game’s narrative limitations. However, it’s also interesting that its storyworld could inspire an established writer in this way, even if the result is more gentle satire than gritty reality. It just reinforces my view of The Oregon Trail as the first really successful ludic narrative — more successful on a narrative level, actually, than Adventure. (Does anyone imagine herself inside Adventure‘s storyworld like so many do inside that of The Oregon Trail?) If I taught junior-high or high-school English, I might have my students each play the game and “novelize” the story that results, even at the risk of reading dozens of reenactments of the Donner Party tragedy, just to see what they came up with.
Will Crowther’s Adventure, Part 2
Crowther’s original Adventure consists of relatively complete implementations of the above-ground section and the first underground level of the complete game that so many would come to know later. It peters out around the “Complex Junction” room, where a sign stands announcing, “CAVE UNDER CONSTRUCTION BEYOND THIS POINT. PROCEED AT OWN RISK.” It’s not kidding; things start to go haywire with some of the room connections at this point, such that navigating in some directions inexplicably returns you to above-ground locations. Beyond the ever-present challenges of navigation, there’s not really that much of a game here. Still, Crowther has laid down the basics of the thousands of text adventures that would follow, and even manages to include a few simple puzzles — and, yes, a maze.
In fact, one could say that the whole of Adventure is really one big maze. By far its biggest challenge is coming to understand and get around in the interconnected nodes (i.e., “rooms”) that make up its world. Even its few simple puzzles revolve around movement: we must deal with the snake to be allowed to progress beyond The Hall of the Mountain King; must find an alternative exit from the cave that will allow us to take the gold with us; etc. This may seem odd, unappealing, perhaps annoying to us when we play the game today — at least, that is, to those of us steeped in the culture of modern IF, with its emphasis on crafting an enjoyable narrative experience for the player. But was Crowther trying to craft a narrative experience at all? I don’t think so, actually.
Crowther is an extremely private person who is not much prone to revisiting the past or discussing his work, so there isn’t much direct evidence as to what he was thinking when he crafted Adventure. We might, however, find some clues in his game’s HELP text:
“I KNOW OF PLACES, ACTIONS, AND THINGS. MOST OF MY VOCABULARY DESCRIBES PLACES AND IS USED TO MOVE YOU THERE. TO MOVE TRY WORDS LIKE FOREST, BUILDING, DOWNSTREAM, ENTER, EAST, WEST, NORTH, SOUTH, UP, OR DOWN. I KNOW ABOUT A FEW SPECIAL OBJECTS, LIKE A BLACK ROD HIDDEN IN THE CAVE. THESE OBJECTS CAN BE MANIPULATED USING ONE OF THE ACTION WORDS THAT I KNOW.”
It’s interesting that Crowther foregrounds the geographical so obviously, and only then goes on to mention the possibility of manipulating just “a few special objects.” As a dedicated hacker, Crowther would almost certainly have come across Hunt the Wumpus. I think there’s a pretty good case to be made that Adventure started as another iteration on Yob’s idea of a “topological computer game,” and quite likely continued largely in that vein in its author’s mind right up until he abandoned its development. It’s very possible, even likely, that compass directions were a fairly late addition, that Crowther initially intended to have the player navigate entirely by working out keywords for getting from place to place, thus making navigation even more of the central chore. (While Dennis Jerz spoke to some who claimed to remember compass directions from the beginning, it’s possible they were misremembering; from reading the source it certainly seems that compass directions were a late — possibly almost a last — addition, perhaps upon realizing just how unworkable keyword navigation was likely to get over the course of a sprawling underground complex populated by dozens of similar rooms.) As a caver, meanwhile, geography would have been constantly on Crowther’s mind, not only as a point of factual interest but literally as a matter of life or death while underground; the in-home teletype connection through which Crowther likely developed Adventure was the same one that he used to enter survey data and construct maps of the real Mammoth Cave for the benefit of other cavers.
How much does it really matter how Crowther conceptualized his game? Perhaps not a lot. Still, it’s worth keeping in mind that expectations of both players and authors were very different back in the day, and that this can explain some things that authors did and players apparently enjoyed which we might find infuriating today. It’s certainly a point I’m likely to revisit again when I look at other historical works. Some scholars have recently advanced the idea that computer games are most of all about the experience of space, even going so far as to call them a form of architecture. It’s an interesting idea, and one that gains a lot of credence when I consider it in the light of these early works of IF. I’m not yet sure how to reconcile that idea with some of my other notions, but it’s more on my radar than ever in light of my experience with Adventure.
Abstractions like that aside, though, there is a certain stately appeal to this early iteration of Adventure which I find hard to explain. Crowther was by neither talent nor inclination a writer, but his terse, matter-of-fact descriptions bear the stamp of someone who knows the environment of which he writes. That gives his game, almost in spite of itself, a certain verisimilitude that would be lacking in many of the more polished efforts that would follow in later years. I want to look at how Woods expanded on this solid kernel next.
TOPS-10 in a Box
Especially since Dennis Jerz’s discovery of the original Adventure source code, there’s been some interest expressed around the IF community and even in academic circles in experiencing PDP-10 software in its original form. The good news is that, thanks to projects like SIMH and archives like bitsavers.org, the tools are out there to recreate history on an everyday PC. The bad news, though, is that getting everything working can be tricky for anyone who didn’t administrate these systems back in the day. Having recently navigated these waters and finally come out with a working system I’m happy with, I thought maybe I could help some others out by offering as close to a one-click installation of a PDP-10 running TOPS-10 as I could manage. So, here it is: TOPS-10 in a Box, being a typical installation circa 1976-1983.
In addition to a complete and healthy TOPS-10 operating system, this distribution also includes FORTRAN and BASIC compilers and — probably of most interest to readers of this blog — Crowther’s original and Woods’s completed Adventure, both in source and executable form. Now you can experience these relics in their original incarnations. The completed 1977 Adventure is particularly interesting to experience on the “real thing,” what with its implementation of “cave hours” and “magic mode” and its strange save system.
In addition, this should serve as a solid TOPS-10 “starter system” onto which you can install games and even other programs. (There’s a nasty rumor that entertainment was not the PDP-10’s prime purpose, although I’m not sure I believe it.) To do that, you’ll have to get to know SIMH and TOPS-10 a bit better, but this system should still give you a leg up in getting started.
The readme file included with the download should tell you everything you need to know to get up and running. Do be aware that you need to acquire one other piece of software, the SIMH emulator itself, and also that this distribution is not small: it’s a 60 MB download in compressed form, and will expand to about 300 MB on your hard drive. So, it’s one for the hardcore time traveler only — but perhaps a few of you who find this sort of window into the past as fascinating as I do can make good use of it.
Will Crowther’s Adventure, Part 1
What remains to be said about Adventure? It has long and rightfully been canonized as the urtext not just of textual interactive fiction but of a whole swathe of modern mainstream videogames. (For example, trace World of Warcraft‘s lineage back through Ultima Online and Richard Bartle’s original MUD and you arrive at Adventure.) It’s certainly received its share of scholarly attention over the years, from Mary Ann Buckles’s groundbreaking 1985 PhD thesis “Interactive Fiction: The Computer Storygame Adventure” to Dennis Jerz’s superb 2007 article for Digital Humanities Quarterly, “Somewhere Nearby is Colossal Cave.” Still, since this blog has kind of turned into a history of early digital narratives without my entirely realizing it, it’s worthwhile to talk about its background. And having recently played it in its original Crowther-authored form as unearthed by Jerz in the course of researching his aforementioned article, I join Jason Dyer in having a few things to say about the experience. Finally, I’d like to make it as painless as possible for you to experience it in that authentic form as well, if you’re interested.
The outline of Adventure‘s history is probably familiar to many reading this, but in a nutshell it goes like this:
Back in 1975 a programmer and spelunker named Will Crowther had just gotten divorced. Missing his children and feeling somewhat at loose ends generally, he started to write a game in his spare time with the vague idea that he could share it with his two daughters, who now lived with their mother and whom he missed desperately. The game, which he named Adventure, combined his three biggest interests at the time: programming, caving, and playing a new tabletop game called Dungeons and Dragons.
How so? Well, the player would explore a geography loosely based on the Bedquilt branch of Kentucky’s Mammoth Cave, a place Crowther had spent years laboriously exploring and mapping; she would encounter treasures and creatures drawn from D&D in the process; and to win she would have to solve intricate puzzles while always maintaining close attention to detail, just like a programmer. Crowther had just invented the world’s first text adventure, in the process prototyping much that remains with the form to this day.
Those are the broad strokes. But let’s back up for a moment. Just who was Will Crowther? Where Wizards Stay Up Late, Katie Hafner and Matthew Lyon’s history of the development of the ARPANET (predecessor to the modern Internet), paints a pretty good picture of Crowther. His eccentricities have become so associated with the hacker mentality that they almost read like items on a checklist today. To wit:
He was almost disturbingly non-verbal, and rarely displayed any affect at all. He refused to dress up for any reason, even visiting the Joint Chiefs of Staff at the Pentagon in sneakers. And “he was a notoriously finicky eater (anything beyond the culinary level of a plain bologna sandwich was a risk), making him an impossible dinner guest or dining companion.” For all that, though, Crowther was a very unusual computer nerd in at least some ways. For one, he loved outdoor adventures, particularly rock climbing and of course caving. As befits an adventurer, he kept himself in excellent shape, in part by hanging by his fingers for hours on the frame of his office door. And most shockingly of all, he “never touched” soda.
Of course, what allowed Crowther to get away with eccentric behavior was the brilliance of his mind. Crowther’s Wikipedia page says as of this writing that, “He is best known as the co-creator of Colossal Cave Adventure.” That’s true enough, but it’s a bit unfair in a way to Crowther that Adventure and caving so dominate the page, for Crowther’s importance in computer history would be assured even had he never created Adventure.
Crowther was an absolutely key player on the tiny team that, beginning in the late-1960s, laid the foundation of the modern Internet. He wrote the software that ran on the Interface Message Processors (IMPs), the set of computers that shunted data around the nascent ARPANET; in other words, he wrote the firmware for the world’s first routers. He was one hell of a programmer, “regarded by his colleagues as being within the top fraction of 1 percent of programmers in the world,” with a particular genius for writing incredibly compact and efficient code, a valuable skill indeed in those days of absurdly limited memory and processing power. If he had a fault, it was that he was more interested in prototyping, in showing that things could work and how, than in doing the hard, often tedious work of polishing and refining that results in a truly finished, production-ready program.
When we add all of this together, we can begin to see how Crowther could have birthed IF in such a complete form almost on a whim… and then abandoned it on another whim when (presumably) a more interesting problem came along.
Hunt the Wumpus, Part 2
To hear Gregory Yob tell it, Hunt the Wumpus was as much inspired by his hatred of the Cartesian grid employed by Hurkle and similar games as it was by anything else. Yob wanted to make a monster-seeking game based on the dodecahedron, his “favorite Platonic solid.” I must say my own interest in geometry is limited enough that it’s hard for me to share Yob’s passion; certainly I lack a “favorite Platonic solid” to compare with Yob’s. I’m more interested in the other innovations Yob deployed on the way to implementing his dodecahedron.
Hunt the Wumpus is the origin point of all those twisty little passages that would be filling so many computer screens and graph-paper pads just a few years after its creation. Its world consists of a grid of twenty rooms, each of which is connected to exactly three other rooms. Some of these rooms have contents, which are randomly placed before each play: bottomless pits that result in instant death, “super bats” that carry the player to another (random) room, and of course the wumpus himself. If the player walks in on him, he has a 75% chance of merely wandering off to another room, but a 25% chance of eating her up right there. The wumpus can be killed only remotely, by firing an arrow from elsewhere into the room that contains him. The game in fact understands just two verbs: “move” and “shoot.” Gameplay, at least if you’re a cautious (not to say callow) sort like me, consists of moving carefully around the storyworld constructing a map of its rooms, connections, and hazards, and finally moving into position to take the kill shot against the poor wumpus. On the terminal, it looks like this:
HUNT THE WUMPUS YOU ARE IN ROOM 20 TUNNELS LEAD TO 13 16 19 SHOOT OR MOVE (S-M)?M WHERE TO?13 I FEEL A DRAFT YOU ARE IN ROOM 13 TUNNELS LEAD TO 12 14 20 SHOOT OR MOVE (S-M)?M WHERE TO?20 YOU ARE IN ROOM 20 TUNNELS LEAD TO 13 16 19 SHOOT OR MOVE (S-M)?
Okay, so it’s not too much to look at. When you play it for the first time, you might end up asking if that’s really all there is. Still, if you give it a decent chance you’ll find a well-constructed little game that can still be engaging, at least for the first few plays as you sort out how it works and how to beat it. From a design perspective, its biggest flaw is perhaps that you can often begin with a configuration like this:
I FEEL A DRAFT YOU ARE IN ROOM 4 TUNNELS LEAD TO 3 5 14
The draft tells you that you are adjacent to a pit; one of those three tunnels, in other words, leads to death. Because you have not yet had a chance to gather any additional information, you are left to rely on blind chance. You must just pick one and hope for the best — hardly a fair situation.
But I’m not so interested in “pure” game design as I am in the history of ludic narrative. From that perspective, Hunt the Wumpus is hugely important in two ways.
First, it represents a radical change in perspective from games like Hurkle. While the player viewed those games from on-high, Wumpus places her in its storyworld. You are there, creeping from room to room in the darkness. Wumpus offers the merest stub of a narrative, but that stub combined with the switch from a third-person to a first-person perspective gives it a very different feel from Hurkle and its companions. Those games feel like abstractions; Wumpus is a much more immersive experience. It wasn’t quite the first game to put its player inside a storyworld — The Oregon Trail, at least, preceded Wumpus by about a year and was possessed of a much more full-bodied narrative in addition — but it’s nevertheless a significant departure from the norm of its time.
Second, and even more importantly, Wumpus is a prototype version of the system of geography that is still with IF today: a set of discrete, self-contained rooms linked together by connectors the player can use to pass from one to another. Compass directions are not yet here, but the rest of the scheme is. Wumpus is all about mapping. The early IF games that would follow were continuing its tradition in being full of those twisty little passages that so frustrate modern players who try to go back to them today. This brings up a point that I’ve only recently started to grasp: the earliest IF was about geography and mapping more so than story or even puzzles. (I want to talk about the original Adventure just a bit after I finish up with Wumpus. I’ll have more to say about this idea then.)
Like The Oregon Trail, all signs point to Hunt the Wumpus having been originally written in HP Time-Shared BASIC. I was able to locate it on tapes preserved by Bob Brown and Michael Gemeny of the HP-2000 Yahoo! Group. Its BASIC code was first published in a mid-1973 issue of the People’s Computer Company magazine, and later appeared in the October, 1975, issue of Creative Computing. The program that appeared there is almost identical to that which we found on the tape, with the only notable difference being some REM and PRINT statements found in the printed version that attribute it to Yob and plug Wumpus 2 and Wumpus 3, two sequels he had written by that time.
Unlike The Oregon Trail, which remained quite firmly under the thumb of MECC and was apparently spread only to educational institutions, Wumpus quickly spawned heaps of ports and adaptations on almost every viable computing platform of its era (and of every era since). By the time it appeared in Creative Computing Yob could write that, “I have reports of Wumpus written in RPG, a listing of one in FORTRAN, a rumor of a system command of ‘to Wumpus‘ on a large corporation’s R&D computer system and have even seen an illustrated version for the Hazeltine CRT terminal!!” It was interesting enough as a game to cross the cultural boundaries that normally kept the cheerful BASIC hippies of PCC and Creative Computing separated from the world of the hardcore institutional hacker. At least by the 1975 release of Unix Version 6 (and quite possibly earlier), Wumpus had been ported to Unix C; a comment in the source cheerfully declares it “stolen from PCC Vol. 2 No. 1.”
Thanks to Bob Brown, you can experience the original version of this relic in its original environment if you’d like, as well as its immediate predecessors Hurkle, Snark, and Mugwump. Here’s what you need to do. (Yes, this is largely the same drill used to access The Oregon Trail on the same system.)
1. Telnet to mickey.publicvm.com. (Telnet, mind you. None of that newfangled SSH!)
2. Slowly alternate CTL-J and CTL-M until you see a “PLEASE LOG IN” message.
3. Enter “HEL-T001,HP2000,1″. Without the quotes, of course — and note that those are zeroes. Oh, and the system isn’t case-sensitive, but for the authentic experience you might want to have your caps lock on.
4. Enter “GET-WUMPUS” for Hunt the Wumpus; “GET-HURKLE” for HURKLE; “GET-SNARK” for Snark; or “GET-MUGWMP” for Mugwump.
5. “LIST” the program if you like, or just “RUN” it.
Have fun!
