I Am Not Myself Today
On characterization in interactive storytelling

For a long time now, I've felt that adventure games are not meeting their considerable potential. My approach to game design revolves around purity: each game should try to figure out what it is that it does well, and then pursue that with a single-minded focus. This attitude has led to many disagreements between me and what seems to be the majority of gamers. Several times a year, some game is widely praised for its exceptionally diverse gameplay, and I crash the party with my insistence that its design is an unfocused mess. "But it's fun!", people protest. Well, I like to think games should aim higher than that. "Fun" is easy. You can get fun by throwing a ball against a wall. So I'm not as concerned with how fun a game is as how good it is, how well it achieves whatever it is it's set out to achieve.

What is it that modern adventure games are achieving? Sure, they can be fun. Puzzles are fun and stories are fun and perception games are fun and exploration is fun, and if you put them all together you're moving between fun activities all the time. But what is the point? Why play an adventure game, when you could get your fun from anywhere else? If it's puzzles you like, you'll be more mentally stimulated by a pure puzzle game. If you're looking for exploration, you'll get a more satisfying (if short) journey in those rare pure exploration games like The Path or Small Worlds. If you want to hunt for objects, there are Hidden Object Games where that's all you do. And if you want a story, there are well-established old Forms for the task. What is it that adventure games have to offer, other than nostalgia for the way computer games used to be made?

Consider this dialogue tree from the 1990 adventure game The Secret of Monkey Island. The context is that Guybrush Threepwood, pirate wannabe, has stolen a statue from the governer of the area. On his way out, he runs into her and sees her face for the first time.

"So, you were just going to borrow it, eh?"

"Uh..."
"Gee..."
"Well..."
"Gosh..."

"Relax, Mr. Threepwood. I know why you're here. Believe me, you're not the first who's tried. Although, I have to admit, not many get as far as you have."

"Er..."
"Um..."
"Golly..."
"Jeepers..."

"My lookout told me of your arrival. I've wanted to meet you ever since I heard your fascinating name. Tell me, Guybrush, why do you want to be a pirate? You don't look like one. Your face is too... sweet."

"Blfft..."
"Grlpyt..."
"Hrdrl..."
"Rldft..."

"I see... Well, you're obviously not in the mood for idle chitchat, are you? I suppose you've got many more exciting things to do. I won't take up any more of your time, Mr. Threepwood."

It's an easy gag, and yet I feel like I can relate to the character of Guybrush Threepwood more in these three nodes than in all the rest of the game. When I play through this dialogue tree, I am struck by a certainty that adventure games have something to offer as a storytelling medium. Think about what the player is going through here. Even before the player has given any input, he is already feeling a bit like Guybrush, caught with his hand in the cookie jar. The fact that there are multiple choices available suggests that there should be a way out, but there isn't one. So the player is confused as to what he should do: is any one of these options actually useful? Then Elaine surprises both Guybrush and the player, by praising rather than chastising him. Again there are four meaningless choices, because Guybrush is so surprised by this that he still can't form a thought. And before he can collect his thoughts, she flirts with him. His brain turns to mush, and the player is given random strings of letters. It is absolutely impossible to communicate. Elaine promptly leaves, and the player is left in a daze. As the player says "This game is unfair!", Guybrush comments: "I really wish I knew how to talk to women." So in a matter of seconds, the player has experienced a fairly complex series of emotions, first-hand!

That's the strength of adventure games. They can put you in the head of another character, and suddenly everything in the world is fresh and new because you're seeing it through someone else's eyes. The player is the character's "free will". The character has a personality, and a history, and everything that a human identity entails, and the player gets to experience it from the inside. That is an experience unique to adventure games among all the storytelling media.

What does this mean in practice? I'm going to try to answer that question from several angles in the posts to come, but it all basically boils down to one simple principle: Only give the player choices which the character would consider. If it doesn't make sense for the character to pick up an object, then don't give the player that ability. If the character really wants something, then restrict the player's choices to ways of getting that. The player should have choices to make, but only in the way that a real person has choices to make. As a human being, you have many years of experience teaching you that some things are acceptable and some things are not. In any given situation there are only a handful of things you will actually consider. Those are the options the player should have. No more, no less.

So whatever the topic of the game, whatever its genre, whatever its plot, and whatever challenges it may pose to the gamer, the very first thing any adventure gamist must ask himself is: "Who is the player character? What makes that person tick?". The rest will follow from there.

Conventional wisdom says that an adventure game's story must compete for attention with its gameplay. Characterization usually comes from dialogue and cutscenes; assuming there's animation and voice acting (as there always is these days), that means we're dealing with techniques inherited from film. And film is a medium built on empathy: you passively watch characters, and try to imagine what they're feeling. The standard gameplay is an entirely different experience, built on direct interaction. The character steps aside, and you solve puzzles and hunt for objects and make choices. The adventure games we're used to playing keep going back and forth between these two entirely separate experiences. I've played adventure games where the story and the gameplay are designed by two different people. It shows, and it's wrong. An adventure game should be one cohesive unit, where being involved in the story is the gameplay.

The key to pulling this off is creative manipulation of the user interface, by which I am referring to the buttons the player will be pressing to interact with the game. This idea runs counter to the usual design philosophies. The interface is seen as a practical tool, nothing more. It gives the player whatever controls he wants/needs over the game world, and beyond that it should stay out of the way. So either the player always sees a bunch of generic "Talk/Walk/Examine/Inventory" controls, or the player is given minimal options for interaction beyond choosing an object on the screen. Either way, the intent is the same: learn how to play, and then never think about it again so that you can get to the meat of the game. The controls are a layer of abstraction that risk pulling the player out of the experience if they get too intrusive.

But the way I see it, playing with the interface is the experience. That level of abstraction is where all the magic happens, so it should be emphasized and expanded rather than hidden away. This may sound counter-intuitive, given that I want the player to feel like he's in the head of the character. Surely the best way to do that is through cinematic immersion, taking away all the game-y bits and pulling you into that pretty world! The problem I have with that line of thought is, the player actually isn't the character, and if he sees the game world directly he's not going to react the way the character would. You can use film tricks like camera angles, soundtracks, acting, etc. to get the player more in the right mood, but those filmic illusions are awfully thin and they'll fall apart as soon as the player is left to his own devices. I see only two reasons why you might ever expect that if you drop the player in the middle of a game world, he'll start acting like the character. The first is if you've designed a game for one very specific person who has exactly the same personality as the character. The other possibility is that the character is so generic that he's not really worth making a game out of. (This is often the case.)

With a more dynamic interface we have another option: we can tell the player how to interpret the game world, by giving him specific choices that frame the situation the way the character would see it.

I'll use a few examples from Gamer Mom (the adventure game I'm writing) to illustrate this principle.

Toward the beginning of the game, the titular mother is talking about her favorite game, World of Warcraft, for anyone who'll listen. The interface is a long list of buttons, each one listing a different thing that is wonderful about World of Warcraft. They can be pressed in any order, and when one is pressed it disappears (leaving the others). Now, I have no idea what any given player's attitude toward WOW will be. He may in fact be an enthusiast, but more likely he's never played the game. And some players will never have even heard of the game. That's irrelevant, because as soon as I frame the situation with the implied question "What order should the game's qualities be praised in?", I have taken the player's personality out of the equation. What matters now is what the character wants, and in trying to satisfy the character's short-term goals the player will come to understand what it's like to be that character from moment to moment.

Here's another example. At a certain point (that the player may or may not reach), the woman's husband remarks that World of Warcraft sounds nice. Now, actually he's just trying to shut her up, but his wife is so starved for attention at this point that she'd like to believe he's interested. There are only two buttons: "He means it." and "He doesn't mean it.". I'm expecting that the player is intelligent enough to understand what's really going on, so the choices should be perceived as "Let's pretend this isn't hopeless." and "Let's just admit the truth.". (I didn't write it that way, because I want to leave some ambiguity as to whether or not she understands her situation.) The two choices take the game in radically different directions, and it's entirely up to the player to decide which way to go. Of course one way is clearly "right" and the other is clearly "wrong", but because these are the only two options provided, the player will (at least for a split second) be on the fence about it just like the character is.

At a certain point she can realize exactly how little he cares about her interests, and the player has many contradictary options on how to react. I will note briefly that the multitude of options intentionally reflects the character's confusion on being confronted with the truth, but that's not what I'd like to focus on at the moment. The most extreme (but still reasonable-sounding) option has her go too far, and she knows she went too far. So all these options disappear, replaced by one little button marked "Apologize". And that button doesn't go away- you can press it over and over and she'll apologize in many different ways, but you never get all those other options back. Now think about what the player will be going through, at this point. A moment ago he had lots of freedom in choosing how to act. Now that's all gone, because of what he chose. I think a tiny bit of guilt would not be out of the question, even though the player couldn't have known how his choice would play out.

As these examples demonstrate, there's quite a lot of emotion you can wring out of a bunch of buttons. Normally you'd expect a button to be a trigger for content, but the buttons are actually the containers of content here. The graphics and dialogue provide context that can convince you you're dealing with a larger world and plot, but all the abstract feelings that give the story its meaning (I have referred to this in the past as the "music" of storytelling.)- that's all in the buttons. How many there are, what's written on them, whether they stay or disappear when you click on them.

And also, how they're placed. Let's say there's a single button to press. This tells the player that there's only one conceivable way to deal with the situation. (Whether that's true or not is irrelevant- it's what the character thinks.) But where is the button? If it's a tiny little button hiding in a corner, it's an uncomfortable step to take. If the button's medium-sized and in the center of the interface area, then it's just a matter-of-fact thing-you-do. And if it's taking up the entire interface area, then it's yelling out "Click me! Click me now! No time to think about it!". The layout of controls can be a useful kind of artistic expression when used properly.

In Gamer Mom the interface is entirely separated from the graphics: following the lead of countless old games, I am placing the interface on the bottom and the graphics and dialogue on top. Because I consider the interface to be of critical importance, I'm giving it a full half of the screen space. I would like to emphasize that while I have only been looking for storytelling techniques that fit this model, it is hardly the only model that might work. Controls could be overlaid on the graphics, the graphics could be a rectangle in the center of the screen with the controls on all sides, and there could be more complex designs where the world and controls are not kept in specific boxes but keep moving around the screen dynamically. And there's no reason buttons would be superior to lists and labels and images and drag-and-drop functionality. I do think it's best to have a pointer of some sort rather than a console-style controller or a keyboard (The controller is too limiting, and the keyboard is not limiting enough.), but beyond that none of what I'm doing is the be-all and end-all of adventure game design, nor is it meant to be. It is a starting point, nothing more.

I imagine with the right interface and a sufficiently creative gamist, any kind of character and emotional progression could be experienced interactively. I can't go that far myself, but I know the first step: the interface needs to be dynamic, it needs to be expressive, and it should never be taken for granted.

We've come to expect 1:1 control of our games. We tell a character to shoot something, he shoots it. We tell a character to move one step left, he does. We tell a character to combine two objects, and immediately the two objects are combined. All this gameplay is very practical -so practical, in fact, that I don't believe it bears any resemblance to the actual experience of being a human being. Humans are messy. We make mistakes. We get distracted. Sometimes we do things without thinking about it, and sometimes we think about things without doing anything about it. To create a character who could conceivably exist in the real world (which is my goal with Gamer Mom), you can't sidestep the messiness of human behavior. So if a player presses a button saying "Stand up" for instance, it doesn't mean the character needs to immediately stand up. All it means is that the character has decided to stand up. Lots of different things can happen after that, depending on circumstance. Conversely, if the character stands up it doesn't mean the player has necessarily pressed a button marked "Stand up", because there might not be any conscious thought behind that action.

As I write this post, I occasionally stand up and pace around the room. There is no conscious thought involved there. If my thought processes were represented by buttons, there might be one saying "Think about the subconscious", and when you press it I just so happen to stand up and pace around the room because that's what I do every time I come across a new thought. (Don't ask me why. I have no idea.) Once I'm pacing, the buttons could be questions I'm asking myself, but there would also be one button marked "Keep typing", and if you pressed that I'd sit down at the computer again. The sitting and standing and pacing are not the relevant actions here, the thought processes are. The movement is just something that naturally goes along with the conscious choices.

Let's say you're eating something. Most likely you're not thinking about every single bite you're taking. It may be a conscious choice to start eating, but past that there's no thinking required. So if you eating something were an adventure game, there would be a button marked "Eat", and once you pressed it it would stay pressed until you clicked on something else that you couldn't do while eating. No "Take another bite" button is required, because as soon as the character decides to eat, all the individual bites go without saying. Now let's change the situation a little bit by imagining that the food is really bad, but for whatever reason the player character is forcing himself to eat it anyway. Then we wouldn't tie one button to many actions, but one action to many buttons!: There would be a button marked "Take a bite", then a "Chew" button and a "Swallow" button. And if we're being cute about it, at the end there might be an optional "Compliment" button, which has the character say "Mmmmm, delicious!".

You could even go beyond these examples. If an action which doesn't require much thought only calls for one button, then an action which doesn't require any thought at all might not need any input at all from the player to be triggered. The character might act on an instinct at a specific point, even though there were no buttons beforehand. (This could cause the player to be frustrated, which can be used for dramatic effect.) And if many buttons imply that the action is hard to do, you could achieve a similar effect by requiring a single button to be pressed several times to trigger an action. Each click before that could visually push the character a tiny bit closer to the action, to indicate to the player that his inputs are in fact being noticed, and that the lack of action is down to the character's hesitation. The gamist can also indicate the severity of the character's fear by making the action's button disappear after a few presses, as though the character loses his nerve entirely when he gets too close to actually doing whatever it is. (On a more cynical note, this is also a useful way for the gamist to avoid having to write a drastically divergent branching path while acting as though the idea was considered.)

By human standards, all these examples so far are very straightforward. The button indicates a clear action, and the character acts accordingly. But we're not always so rational. If there's a button marked "Get book", but that book is in a different room, pressing the button might just lead the character to that room, where the player finds the buttons "Why did I come here?" (having no effect when pressed) and "Go back". It's so perfectly common to forget things; why shouldn't it be common in adventure games too? Alternately, the character could get distracted by something else along the way to the room with the book, leading him on a tangent that he never gets back to the book from. If nothing else, it's a tremendously useful little plot device: you have the character going to deal with something tedious, that unexpectedly leads him to a place where something interesting is going on, and the chore is swiftly forgotten. The action that the clicked button suggests doesn't actually need to ever happen.

A thinking person is always going to have more options than just actions to choose from. There are all sorts of other decisions to make: what opinions to form, what attitudes to display, what ideas to keep in mind for later. These buttons would generally be marked with statements rather than commands: "I don't like this person.", "I'd better not waste any time.", "Interesting.". Clicking on one of these buttons does not make its statement true; if the statement appeared on a button, then the character is already thinking it. Rather, the player in clicking on a statement is giving the character implicit permission to act on that thought/feeling in the future. In certain cases it'll make sense for the character to do an action much later in the game without any input from the player, just because the player gave that permission here. But more often, it will subtly add or subtract buttons for the player to choose from and/or slightly change the actions triggered when the regular buttons are pressed. It is not always essential to include buttons to pay off an earlier button press, but often it makes sense.

Buttons can also have immediate effects even if they are not clearly marked. A polite person will act on the button "I don't like this person." by having the option to end a conversation with that person a little earlier, but a rude person may immediately say "You know, I don't like you very much.". Obscuring the end result from the player can reflect that the result is unclear to the character- he doesn't necessarily know where any thought he follows is going to lead him. A button called "Disagree" may instead (in specific contexts) be marked "That's not true.", and the difference in how the player experiences it is that he's unsure whether the button is an immediate command or just something to keep in mind for later. I find that foggier actions make for a more authentic experience.

One thing which I'm absolutely not going to do in Gamer Mom is to present dialogue options exactly how they are spoken. I am referring to the standard way of handling dialogue trees: you have a list of complete sentences, and when you choose one the character repeats it and receives a concise response from whoever he's talking to. Playing through dialogues like that, I always get the sense that my character isn't being honest and human, he's just reciting a script I'm feeding him. That's why I'd rather keep choices to things like "Reluctantly accept", "Go back to the other topic", etc. If you see the entire statement before it's made, then the character knows exactly what he's going to say before he starts. Who talks like that? Even if a character thought he knew exactly what he was about to say, I'd probably fill his statement with "um"s and "like"s and "y'know"s, and I'd have fun getting the other character to interrupt his speech in the middle with questions and all sorts of messiness like that. Humans do not act like computer programs, outputting data efficiently. So if you want a computer program to be about humans, you've got to add in all that messiness that we never think about but always do.

The next time you catch yourself messing some little thing up, or doing something you didn't exactly intend, think about how that behavior might work in an adventure game. You might be surprised by how easily the experience of being you translates into gameplay.

From a certain perspective, everything I've said thus far is common sense. I want to encourage the idea that the adventure game in its pure state is a common-sense kind of medium. After all, is it not a universal idea to wonder what it would be like to be other people for a little while? It seems perfectly natural for a game to try and sate that curiosity. But while this idea is simple, its implementation is not. It is a good start to decide on a consistent characterization and convert that characterization into an interface. But still what you're left with is just an illusion: that the player is being given control of a real person with a personality and a life. Illusions are flimsy. At any point the player might remember he's playing a meticulously-scripted game, and stop caring about the character.

Thankfully, the player is probably on our side. There is a tendency to recognize patterns and human behaviors in all things, so the player will usually make an effort to fill in the gaps in the game's reality in his mind. When we offer the player an abstract reality made of buttons and images, we can realistically expect him to imagine that there is a "real world" on the other side of those buttons, just because the player is human. Don't take it personally, it's just how we're wired. A human brain, upon seeing patterns that seem like human behaviors, immediately assumes there's another human on the other side. In the case of a (single-player) adventure game, there's obviously no person on the other end. So many players will automatically assume that there's some "artificial intelligence" program in the code, with a personality and something like human thought processes. (There is, of course, no such thing going on.) Those who understand the techniques used will reject such ideas, but they might have similar feelings going on under their rational skepticism.

At a certain point, an adventure can stop seeming like a computer program and start feeling like it's just people being people. At that point, the player is hooked. Whatever happens in the game after that is real and has the corresponding weight. The gamist should attempt to bring the player to that point of irrational belief, and keep him there for as long as possible. While the player believes, the adventure has a self-evident value. But as soon as he does not, the game becomes a curious novelty and nothing more.

The first step, as I've mentioned, is to follow the principles I laid out in the last three posts. Without a dynamic, character-driven interface, there's nothing to fool the player with. Beyond that, it's mostly a matter of finding a realistic degree of chaos. Remember, what the gamist is fighting is the perception of a straightforward computer program. "This isn't a game, it's a person you're playing!". So anything which seems too computerized and logical damages the illusion, reminding the player that he's actually just clicking on buttons on a screen. And contrariwise, things which don't make sense and aren't expected add to the illusion. This may seem backwards: why prevent the player from finding patterns? But the fact of the matter is, the real world is never perfectly straightforward. It makes sense, yes, but you need to try to make sense of it. Something that's perfect and simple is perceived as being artificial, whereas certain kinds of imperfections and complexity seem to indicate an underlying reality.

The player should never get the sense that the non-player characters exist for his benefit. A character who stands around dispensing information on request is a massive gaping hole in the illusion. The same goes for NPCs who perfectly repeat their statements or actions indefinitely. But when an NPC acts in a way that does not serve the game in a clear way, that enhances the illusion. Including characters who don't want to be bothered and won't help you get anything is clearly not an efficient way to write a game, but that's exactly why it'll work. An NPC who interrupts you or hinders your progress in some small way is an even better method. Reality means messiness. You can't create the illusion of reality if you're not willing to be wasteful sometimes. I'll deal with this idea more in the next post.

It is not exactly wrong to use voice acting and animations to increase realism, but it is problematic. When you make an animated movie, you get the timing just right so that everything looks natural. But in a game, the player has a lot of control over the timing. Also, any given node may be arrived at from several different directions due to the necessity of tying branching paths together. So the animators and actors can't ever know the exact context their work will be seen in. The more detail you try to put into those features, the more those little imperfections will be noticed. These are not huge problems, because most players will choose to overlook the problem in order to be better fooled. But personally, I'm choosing to rely on static images and text wherever possible. The player can imagine a world between the images, and I'm not going to risk breaking that with specific transitions. But the cinematic approach has its merits, and I wouldn't begrudge a gamist who chose to go that route.

When I first started writing the script for Gamer Mom, the concept of a realistic character-based adventure seemed so obvious that I wondered why no one had done it. I quickly found my answer. You see, human beings are complicated. To mimic them convincingly amounts to countless hours of hard work. There's no way around it. You can cheat sometimes, to let yourself have less work, but it feels like a cheat to the player. It hurts the reality of the experience. If you provide only one option when the characterization realistically calls for three, you might as well be making a movie rather than an adventure because you're wasting the potential of the medium. (I've played way too many adventures where I wonder if the story was envisioned as a game or a film.) If you're doing the job properly, the game just gets more and more complex exponentially, until it gets completely out of hand. You can tie branches together sometimes, but if you do it too often the player (rightly) gets the impression that nothing he does matters. So you just have to follow the ideas wherever they go, regardless of how much work this leaves you with. Why hasn't a realistic character-driven adventure been made before? Because it's frigging hard, that's why! It's a struggle just to release a slightly-interactive film with "Where's Waldo?" segments, let alone a true adventure game.

So what's usually done is to limit the player's options. As much as gamists like to talk about branching paths and giving the player control of the story's progression, every player who finishes a game will usually be seeing at least 80% of what's been put in. Interactivity often has no impact whatsoever, or only controls what order you experience things in. Now, a good or even great adventure game can be made under these restrictions. But it needs to be a part of the premise of the game. Gamists should not expect their players to be stupid- if the player has no control over the plot, it is only a matter of time until the player feels like he has no control over the plot. And that then becomes an integral part of the story he's experiencing, whether the gamist intended it or not. So the gamist should intend it. By all means, tell a story about a character who is not in control. Tell a story about a person who has things happen to him rather than ever being proactive. The problem is only when there is a situation where the character realistically ought to do something, but won't. A lack of interactivity breaks the reality of the game, unless the reality of the story is built around hopelessness from the start. The story can be about someone with minimal intelligence or willpower being led around, or it can be about a person obsessed with routines, or it can be about mundane things that don't matter. But as soon as the story gets more ambitious than that in its characterization, the writer has himself a problem.

I can understand why such an approach is taken, of course. If you put in work, you want for that work to be seen. You don't want a player to only see 10% of what you've done. But you should want that. If the player only gets 10% of the content, he's earned that content. He's gotten the 10% that he chose himself, and feels like he's been the character for however long you've let him play. He'll have regrets over what he didn't do, and pride in what he did, and everything which you'd realistically have if you stepped into another person's shoes for a day. But if you show everything without being prompted to, then very little of that is going to feel earned. The player will feel that he has watched a good story, perhaps, but not that he experienced one. A good interactive storyteller (in any Form) needs to be willing to let his work go unseen, or else the experience will feel impersonal.

A better way to deal with the exponential-scale problem is to limit games to several minutes in length. I have finished the script for Gamer Mom all by myself, because I specifically chose a very short and simple story as its subject. Granted, it took me over five months, but if I had approached it as a full-time job I'm sure I would have been done much sooner. The script does not skimp on depth, because no matter how you play it won't go for very long. A player could load the game, play for two minutes, reach an ending, and leave. He will have seen maybe 1% of the game, but this is an acceptable scenario for me. Whatever ending he reaches is his ending. The story is complete no matter how he's played, it just won't be the same story that another person playing the game will get. But each player, oblivious to the depth of the game, will feel that they've just experienced a real-world situation, because that's how the real world is. You don't see everything. You get to one ending, and then you move on. There are all sorts of nuances to the characters and plot of Gamer Mom which I fully expect that only a tiny fraction of players will ever see, including an entire extra scene at the end. This is by design. If you reach those branches, they'll mean more to you.

Of course, when you embrace depth there's a problem with the workload for the artists. I'm continually concerned that Kyler will back out from drawing the game, worried about how much effort it'll take to turn this 34-page script -with around 600 nodes, I'd estimate- into graphics. I've assured him that he can reuse images as much as he sees fit, but it's still a lot of work. When I'm making bigger projects (with less depth, of course- there will always be a trade-off) I may have a budget, and I'll be able to pay artists for their time and effort. But Kyler is entirely volunteering, like I do in acting, and as much as I trust Kyler I couldn't exactly blame him for getting scared off.

Let's imagine an adventure gamist who knew what he was doing running a massive adventure game company, with the resources to combine length and depth. Another option then presents itself. If one writer cannot do justice to a reasonably-ambitious story, maybe fifty writers can! I envision a hierarchy. One or two people at the top ("editors") come up with the story in broad strokes, and continue to shape it throughout the writing process. Then there are lead writers, who have regular meetings to share their ideas and progress and make sure that everyone is on the same page. They write the critical scenes where the player makes major decisions, and then follow those decisions to their logical conclusions from scene to scene. The lead writers would work in collaboration with each other whenever they reach scenes dealing with plot threads developed by more than one writer. The editors would read through all of the lead writers' incomplete work (written strictly in the playing order, to minimize the risk of having to throw out large quantities of script when changes are made), and decide which branches can serve the larger story and which are dead ends. They then assign innumerable minor scenes to junior writers to connect the major scenes together, with instructions to either come up with reasonable ways of tying the branches into other branches, or follow through on a thought without allowing too many new branches to be created.

In this model, the presentation of the story must be limited. The more detailed the graphics and sound, the more man-hours are required for each line of script. (And by this point the page count would number in the thousands!) With photorealistic worlds and motion-captured animations and detailed sound effects and voice acting and whatever other frills we see in modern games, the required team size (and budget!) would quickly outstrip Hollywood movies. So the presentation must be abstract, and simple. A certain amount of prose is acceptable; animation is not. Sound and music should be kept to a bare minimum. Ideally there should be writers with an understanding of design principles, whose entire job is to read through the script, and write a more detailed script with complete directions for how it should be presented.

I am not qualified to propose a business model for this approach to adventures. But I have faith that some crafty capitalist will figure it out, and someday we'll have adventure games which live up to all the promise of the Form.