ePilgrim’s Progress

My Experiences with the Erasmatron, A New Tool for Creating Electronic Fiction

by Laura J. Mixon

Fiction's future in the electronic medium is on many SFWAns' minds these days, with the explosion of popular interest in computers and the Internet. In our works we fantasize about virtual worlds; we wrestle with publishers over electronic rights; we wonder how to go about selling game rights (we toy with the remote hope that our latest book might be the next Doom or Myst, if only we could break into the market); we've all heard about Sony Bookmans and the like, devices that will supposedly replace paper books someday. On our darker days, we wonder if we'll be outstripped, replaced by some newer form of entertainment that will leave us, like the minstrels and oral storytellers and epic poets of old, in the dustbin of outmoded entertainment.

Over the past couple years I've been a pilgrim in the brave new world of electronic entertainment. In mid-1996, Chris Crawford approached me with an offer to create a prototype work that would showcase some software he's been working on.

Let me pause to give you some background on Crawford. He has been designing computer games since the late seventies, when the computer gaming industry was pre-fetal -- merely a gleam in a few crazy programmers' eyes -- and his reputation for creating intelligent and offbeat games has won him all sorts of commercial and critical success. His game "Eastern Front 1941," for instance, sold 60,000 units in 1981, and "Balance of Power," released in 1985, sold 240,000; bear in mind how few people owned personal computers back then and you'll have a better idea of just how big a hit these games were. His games have also won critical attention for their highly innovative and unique designs.

But his influence goes far beyond a contribution to innovative game design. He founded the Computer Game Developers' Conference, the annual gathering of game designers and programmers, and published the professional designers' journal Interactive Entertainment Design. He lectures at conferences and universities around the world and has been profiled in many places, including The New York Times Magazine and Newsweek. Though the analogy isn't exact, you might say that Crawford is the computer gaming industry's answer to Damon Knight. In short, Crawford has been a seminal influence in the field of computer game design.

For the past five years, he has been developing a tool for professional storytellers. This tool, the Erasmatron (named after Crawford's all-time hero, Desiderius Erasmus, the sixteenth century writer), allows writers to create a totally new kind of electronic, "interactive fiction." A new narrative form, tailored to the strengths of the computer, not merely transplanted from paper.

It's a big claim and deserves a hard, skeptical look. But I have come to believe that he's really onto something, and I want to share my findings with my fellow SFWAns.


Why a New Form?

The first question that needs to be asked, perhaps, is why do we even need electronic, interactive fiction? What's wrong with good, old, tried-and-true books?

Well for one thing, tens of millions of people are on the Web now. Conservative projections are that twenty percent of American adults will be on-line by the year 2000 (according to some estimates, twenty percent of US adults already are on-line). The electronic entertainment industry, with its huge audience and instant accessibility, has a market potential that makes the movie industry pea-sized by comparison. People poised to tap that market with the right product are going to reach millions with their works -- and make a killing.

But to tap that market, we must play to the strengths of the computer, not merely try to uproot and transplant techniques that work as well, or better, in print.

And we all know time doesn't stand still. Imagine Homer and his contemporaries muttering discontentedly over the idea of putting their epics down on paper. Ye gads! Heresy! It'd put us all out of business! The fact is, the human race needs storytellers as much as it ever did; it's merely the medium and mode of the telling that have shifted.

Computers are begging to be used to create something new, something wonderful and different that will allow us to speak to the human condition in ways we can't approach with traditional prose. If we don't do it, someone else certainly will.

The State of the Art

The industry that comes closest to creating "interactive electronic fiction," whatever that is, is arguably the computer gaming industry. But as anyone who plays computer games knows, games are not stories. They're long on jazzy graphics and special effects and notoriously short on heart. The most popular games on the market comprise either twitch games -- shoot-'em-ups and drive or flight simulators that give your hand-eye coordination and your cardiovascular system a workout -- or puzzle games that take your logical and deductive skills out for a spin. Both types (especially the former) can be quite compelling if you thrive on adrenaline or cerebral highs, but they have all the emotional depth and thematic richness of your average halibut on ice in a fish market.

There are also the educational products that have more emotional content but which are only minimally interactive, such as Broderbund's "Living Books," that let you fiddle around inside a given screen on request, but the overall arc takes you through a preset path. They're not nearly as popular, and for good reason -- they don't give the user any real control over the story's outcome.

The fact is, the nature of computer entertainment is controlled by the priorities of the people who create the products. The programmers. People who, to a large degree, have spent their lives since adolescence immersed in computers and program code, most of whom are as uninterested in creating anything with story content and emotional range as a group of people can get. And those who are interested in more innovative efforts are trapped by the narrow, niche-marketing expectations of their publishers.

And so the industry has evolved to its natural endpoint: a large, complex mechanism for producing games with sexy graphics, high-quality 3D imagery, and surround-you-with-sound, and meanwhile ignoring the very essence of what makes fiction so compelling: human beings, wrestling with real problems. Until a product comes along that infuses the compulsive playability of high interactivity with the rich, fulfilling sense of character and theme at which the non-interactive media such as print excel, the computer entertainment field will remain a niche market for young, white, male techno-geeks.

Don't get me wrong. I've got nothing against young, white, male techno-geeks; I married one, and except for the male part I fit the profile to some degree, myself. But most people don't. So a computer product that will appeal to those millions of people who don't fit the profile of your typical computer geek must be something different than the current crop of computer games.

Interactive fiction will be something different.

The Nature of the Beast

The jury is still out on what interactive fiction will be. The field of computer entertainment itself is still very young, less than fifteen years old, and already any number of structures have been created, used, toyed with, and discarded. To make the question easier I propose to break the phrase down into its components and examine each separately.

Let's start with the easiest term. "Electronic" means stored, displayed, and/or transmitted digitally -- using our computers. It might be accessed and downloaded from the Web, or we might purchase a CD-ROM or (how archaic) even a floppy disk.

As for "fiction" -- in its broadest sense, the American Heritage Dictionary defines it as "a work of the imagination." One of my Clarion instructors defined it once as a bunch of lies strung together in such a way as to reveal a deeper truth. This latter in particular rings very true to me.

And lastly, "interactive." Spend ten minutes with a group of computer game designers and you'll hear this word bandied about at least a dozen times. Turning again to my American Heritage Dictionary, I find the word referred to as the adjective form of the verb "interact:" "to act on each other."

Here we've stumbled upon a critically important concept. Interactivity is the single thing that computers offer that traditional media such as movies, theatre, and prose can't (though perhaps theatre and oral storytelling come closest, with the interaction that occurs between the actors and the audience).

But, "to act on each other;" what does that mean, in computer terms? What is the nature of an interaction between a human and a computer?

Listen, Think, and Speak

Crawford likens an interaction to a conversation, in which each participant does three things: listens, thinks about what the other has said, and speaks. In order to have a truly satisfying interaction, each party must be able to do all three well -- someone who speaks well but neither listens to nor thinks about what the other participant says is not interacting very effectively.

This analogy holds true for computers. To be truly interactive, an electronic story must not only speak to the human engaged with it -- which is what print fiction does so well -- it must also listen, and think about what the human's responses mean. It must in some manner be aware of, and must give control over the story's outcome, to the human participant.

What this boils down to is that the writer of electronic, interactive fiction must visualize the goal not as a finished product, such as a book -- but as a changeable entity, one that in some manner is aware of and responsive to the reader. And to create a computer-based entity that responds to the reader, you must create an executable program that when activated presents believable characters doing believable things: a program in which the reader is constantly presented with relevant, meaningful choices that affect the outcome of events in dramatically important ways.

For many of us, the idea of creating a computer program sounds pretty damned intimidating. And even those of us who aren't intimidated figure, what's the point? We're writers, not programmers (at least, most of us aren't). The idea of spending years learning how to write computer code -- climbing to that level of abstraction and, yes, tedium -- in order to create this new brand of interactive fiction, doesn't thrill and excite.

On the other hand, the medium's potential has tantalized me for years. I've been bored and even irritated by the creative shortcomings of the usual run of computer games, but I sense that the medium can be used to do so much more than it currently is.

Which is why, when I heard about Crawford's electronic story scripting software, the Erasmatron, I was intrigued.

What The Erasmatron Is

The Erasmatron is not really like anything else out there, so bear in mind that all of my comparisons are lies (perhaps if I tell you enough of them, they'll reveal a deeper truth).

The product it's most like is perhaps Macromedia Director, in that both are "scripting" tools for professionals, non-programmers, that allow you to create executable, interactive files. But the similarity ends there. Whereas Director is a tool for creating and manipulating images and sounds, the Erasmatron, at its heart, is a storyteller.

The Erasmatron provides you the ability to create executable, interactive story-programs, while insulating you from the arcane complexities and general finickiness of traditional programming. It's a beautifully clever system that puts a tremendous amount of power in your hands with considerably less investment on your part than having to learn, say, C++ or Visual Basic (programming languages often used by game designers).

Your job as storybuilder is to design and assemble the story components, which consist of characters whose looks, personalities, and relationships you define; the stages on which they interact; the props they can use to interact with; and most importantly, the web of "verbs" that define what they can say and do to each other. The Erasmatron then bundles that up with the storytelling engine, which triggers the story and acts out the script you've prepared for the characters.


The End Product

The Erasmatron, like a word processor, is simply a tool used by the writer to create an end product -- a "storyworld" -- which may then be marketed via the Web or sold to an electronic media publisher. I'm developing the Erasmatron's first prototype storyworld, entitled "Shattertown Sky," which will be made available when the software is released, to display some of the Erasmatron's features and capabilities.

The storyworld's "front end" (i.e., what the reader/player will see) is still in development, so the exact look of "SS" at runtime is still an unknown for me, as of this writing. I've seen an earlier, demonstration storyworld and I do know that the end product will have four major elements. Its central feature is a large face display showing the character currently interacting with you (characters' faces are assembled by the writer, using palettes of faces, hairdos, and costumes provided by the Erasmatron's Face Editor tool; characters' expressions change with each interaction, as specified by the writer from a list of over a hundred possible facial expressions). In addition, a menu bar at the top of the screen provides you with background and help, the dialog and narrative are provided beside the face, and below the face and narrative is a list of your possible responses as reader/player.

The first-generation storyworlds built using the Erasmatron are mostly text-based -- less graphics-intensive than the current crop of computer games. There are some graphics: non-player characters' faces show changes of expression in response to the player's actions, and a certain amount of animation of the faces might be done, but there is no sound; all dialog and physical interactions are presented textually. I anticipate that, if this software and its products take off, the front end will eventually be souped up, either by Crawford and/or by someone else. For now, though, the storyworld's greatest strength lies in its guts and heart and mind, not on its skin.

Where You Start

To create the body of your storyworld, you use a set of scripting tools provided by the Erasmatron. The software features point-and-click editing, so scripting your storyworld proceeds quickly once you get the lay of the land. There are a lot of important techniques you must master, and, because this is a highly sophisticated tool for professionals, a whopping lot of tools and features you must learn the uses of. The good news is that the design is very smooth and intuitive -- the software is surprisingly easy to play around with. It's easy to experiment. Extensive tutorials, as well as built-in and on-line help, are in development for release with the product. And Crawford's playfulness shines through the design; you'll find delightful evidence sprinkled throughout. (One of my personal favorites is that you get to "shoot the programmer!" if you bump up against one of the software's limits.)

But don't misunderstand me; this tool takes time to learn. I've been using the product intensively since September 1996, and there are still features I haven't used. The point is that you can get started at a very basic level with some initial knowledge, and as your need to do different things with your characters and their actions grows, you'll have access to the more sophisticated features and the associated on-line help.

Conceptually, designing a storyworld has a bit in common with designing a role-playing game. You assign personality traits to your non-player and player characters and define their initial relationships, you create the milieu in which they interact, and you design the rules governing their interactions. You also write the dialog and narrative for each potential event using the TinkertoyText Editor, a clever tool that presents multiple versions of a given event, depending on who's doing it or relating it the reader/player. At runtime, the storytelling engine acts as your avatar, guiding the human player and computer-generated characters through the web of potential events you've created.

I should note that, for the most part, you don't specify the actions each character takes or when; instead, you define potential actions that any character might take, and you specify under what circumstances they might take those actions.

For instance, you might create a verb called "AttemptKiss." At runtime, then, John might try to kiss Myrtle or Joanne; Mark might try to kiss Ann, Marie, or Helen; Helen could try to kiss Val or Pete or even Myrtle; and so forth, depending on the constraints you impose. The possible responses the kissee might make could range from "SlapFace" to "PeckOnLips" to "RedHotKiss," and the kissee's actual response could depend, for instance, on his or her libido, affection for the kiss-attempter, timidity, and pride.

Thus, the storyworld you create is made up of these many verbs and their components, comprising threads and loops of interaction that move out from the storyworld's center (i.e., the beginning) to its edges (i.e., the multiple endings), depending on the characters' -- particularly the reader/player's -- choices.

Be prepared to haul out and dust off your high school algebra; to direct the characters' actions you must create "inclination equations" that the computer uses to figure out what a character will do. For the most part, it's pretty basic stuff. For instance, using the "AttemptKiss" example above, if John attempts to kiss Myrtle, the inclination equation for her possible responses might look something like this:

1) Inclination[SlapFace] <= Pride[Myrtle] - Charisma[John]
2) Inclination[PeckOnLips] <= Affection[Myrtle,John] + Timidity[Myrtle]
3) Inclination[RedHotKiss] <= Libido[Myrtle] + Arousal[Myrtle]*
1) Myrtle's inclination to slap John's face is equal to her pride minus John's charisma.
2) Myrtle's inclination to give John a tepid peck on the lips is equal to her affection for John plus her timidity.
3) Myrtle's inclination to give John a red-hot kiss is equal to her libido plus her level of arousal (a mood).

Personality Traits, Relationships, and Attributes

The Erasmatron allows for a cast of up to thirty characters, each with twenty-eight built-in personality traits and attributes, eight relationships (such as affection, trust, and submission), and four moods. There's room for you to create more attributes, traits, relationships, and moods tailored to your own storyworld's needs. You can also assign up to eight attributes each to things and stages (e.g., for a thing, you might assign characteristics such as hardness, portability, whether it's a projectile weapon, whether it's a family heirloom, etc.; for a stage, you might assign characteristics such as whether it's indoors or outdoors, its ease-of-traversability, aridity, ornateness of decor, pollen count, amount of air pressure, extent of structural damage, etc.). These characteristics can then be modified by events, and can also be used to affect characters' actions. The software also features PlotPoints, a special set of functions you may use to set up time-driven events. PlotPoints function independently of (or interdependently with) the action-reaction threads of character interplay, and thus allow you to schedule key events and provide a dramatic arc.

This degree of sophistication, especially with regard to characters' personalities and relationships, is well beyond anything else currently out there.

The Storyworld Structure

The most difficult part of learning to use the Erasmatron, for me, has been learning to think differently about what I'm trying to accomplish. There are some fundamental differences between an effective storyworld and an effective print story, and at first glance the two seem to conflict. Tossing out old paradigms has been a major struggle for me, but it is absolutely necessary to rethink your approach when designing a storyworld.

The two most important differences between an electronic storyworld and a print story are, first, that the storyworld should not be linear -- i.e., while the individual links in a thread are causal, the different threads must not be overly causal, and, except for PlotPoints, all actions should be combinable in different arrangements. And second, the storyworld should not have a single, set outcome. In other words, to make the reader/player's experience meaningful, the storyworld must have multiple pathways to various endings that are dramatically different (if their choices don't affect events, your audience is going to wonder why they have a choice at all -- why, in fact, they're not just reading a book or watching a movie). So, instead of thinking in terms of a single chain of events leading to the Big Climax, you must think of your storyworld as a galaxy of possible events winding past, through, and around each other. Thus, with each playing the reader/player will experience a different story, by making different choices.

To accomplish this you must think of many different takes on a given theme. For instance, "Romeo and Juliet" is one take on the theme of romantic love in conflict with political and familial forces. To have an effective storyworld, you need to think of several stories around that theme, not just the one we're familiar with from the play. Several main threads must weave through, past, and around each other in ways that resonate with your story's central theme.

Once you shake yourself out of the freight-train, single-chain-of-events approach, this is actually exciting -- all those dead ends, snake hands, and off-on-a-tangent branches that you have to prune out when you're working on print fiction are actually a good thing to have in your storyworld.

Multiple-Path Creation

One technique that has worked well for me is to identify the player as the protagonist and imagine that the story has a number of different protagonists (as, in fact, it will), and ponder how their differences might affect the story's arc. Another is to focus on the theme itself, try to articulate clearly to yourself what it's about, and then ponder what different kinds of stories might work to illustrate that theme in various ways.

I fully expect others will use the Erasmatron to create storyworlds of such diverse vision and creativity that I can't begin to grasp them now, and my prototype will merely be a rickety, crude first step into this new medium. And the Erasmatron itself is young, and will undoubtedly be improved upon and expanded as feedback from storyworld writers enables Crawford to polish his storybuilding tool.

But I have become convinced that, with the Erasmatron, Crawford is handing us a tremendous opportunity: a powerful tool that we can use to transport the strengths of fiction to a whole new realm -- and perhaps even revolutionize the computer entertainment world (hey, don't forget, Crawford has done it before).


The Release

The Mac version of the Erasmatron will be released in August, 1997, with the PC release scheduled for January, 1998. To use the software, you'll need the following:

MacintoshIBM or Clone PC
15-inch monitor
15-inch monitor
Power PC (i.e., not a 680X0)
a fast 486
16 MB of RAM
16 MB of RAM

The Erasmatron will sell for about $200. The storytelling engine, which is bundled up with your storyworld data and runs it, will be royalty-free -- in other words, you won't be required to pay a royalty to Erasmatazz to sell or distribute your storyworld. The software may be purchased and downloaded from their Web page at http://www.erasmatazz.com/. They plan to offer a page for authors to upload their works for distribution, either as freeware, shareware, or demoware.
-The End-