#include: the usual disclaimers about long posts and navel-gazing.
My ifcomp entry, Out of Scope, ranked poorly in the competition this year, and this post is aimed at restitution via revision. I want to soften its sharpest edges for a post-comp release and contemplate what a theoretical second entry would look like, having learned a lot from this year’s experience. There were several recurrent and thought-provoking themes in the feedback, for which I continue to be grateful:
Several bugs prevented players from progressing, and the worst of these were in the fourth chapter, of nine, where a couple of uncaught permutations left the game unresponsive or in a dead-end state that required a reload to reset it. By the third week of the competition, reviews started to focus on the game’s content and user interface again, meaning that I think I fixed the biggest bugs, but by then plenty of experiences had been spoilt.
I don’t have testers with much stamina in my personal life, but if I compete again, I’ll try to scout some from this community before release. I started too late (June 23) and was too shy to try it this year, but everyone I have interacted with here has been the model of generosity. I can just about conceive of inconveniencing others on my behalf if I can offer a trade in return, as the competition guidelines suggest.
Most of the bugs on this project came from its homegrown systems, which of course the competition guidelines warn about. Having control over my own grammar of interaction is important to me, though. Dreaming about novel ways to experience text is a big part of the fun, so it would feel like a sacrifice to switch to an established framework like Twine. However, there is some fallacious thinking at work, there: in order justify the systems work, I’m tempted to double down on potentially bad ideas.
Speaking of… the most criticized aspect of “Out of Scope” was the user interface, but it was also the incipit for the project. I find reasoning about space to be a fairly taxing part of playing IF myself. Many good counterexamples avoid this by making space less significant to decision-making, by providing a map or expecting users to make their own. Could there be a natural way to imply spatial relationships without having to produce art assets or dedicate a lot of boilerplate description to it? 3D video games communicate space at a glance, but environment modeling is labor intensive and these games struggle to represent surreal or metaphorical spaces effectively. Maybe a 3D space where the proximity of paragraphs to one another and spatial audio forms an imaginary scene would work. (It had to be 3D; 2D proved too crowded with even a sparse set of symbols.)
Well, that was the thesis. Even after reading all the complaints, I’m still not sure if this is a good idea that needs refinement, or if it’s a moonshot whose failure points to energy better spent elsewhere.
The first and most embarrassing complaint was that it was unusable on a trackpad. It’s embarrassing because I just never thought about this as a problem. I occasionally developed the game using a trackpad, so it should have occurred to me that a prolonged period of clicking and dragging on one would be painful.
Two different remedies were suggested in the reviews: keyboard control, and a camera-follows-mouse design that would be familiar from first-person POV game genres. For the last release, I’ve elected to add keyboard controls: either the arrow keys or WASD will point the camera at the text that is closest in that direction, and the space bar or enter key will engage with it. The directional keyboard inputs can also be used to control the choices (accept or reject). I chose this over camera-follows-mouse because that method would require a modal switch between looking around and choosing options or using other interface elements; it would require toggling on and off the mouse cursor; it would make playing the game windowed harder; and because I would have to add a visual indicator (like a crosshairs or animation) for which text would receive focus when the pointer was pressed. I’ve been totally wrong before, but when testing it I think the keyboard input method feels good, and it can be used to play the whole game without any pointer input at all.
The second complaint is the one I’m most worried about: some people found it dizzying or nauseating to rotate the camera. This worries me because I’m not confident about how to compensate for it. I’ve tentatively considered a few options:
I could add a horizon line, like a gradient from the center of the screen that will give more gyroscopic reference, especially when the screen is sparsely occupied or unoccupied. However, I believe that this will also imply a wide open space, which would be dissonant when representing interiors. Someone else suggested a radial gradient that emanates from each text bubble, which would create a color-mapping between parts of the scene that are closer to a point of interest or farther away. I don’t know – I think it would imply a literal glow or light in the imaginary space.
I could increase the field of view. Right now word bubbles whip by quickly when rotating the camera because the visible field of view is quite narrow. Widening it would result in there being less visible motion when the camera moves and it would take less input to completely rotate one’s perspective. The cost is that it diminishes the visible effect of space on the appearance of the text – bubbles that are far away from one another will not look as differently sized as they used to. The visible influence of distance is part of the reason for using a first-person analogy in the first place, so the cost would be steep, but perhaps there is a better sweet spot balanced between the extremes.
I could add more visual reference points, like a map or a “height line” that traces the gap between a text bubble and the ground. However, I’m wary of the extreme potency that graphic symbols and patterns have for distracting the eye from text, and I’m also reticent to commit to a system that will require me to author art alongside writing. One beauty of interactive fiction, compared to the classic video adventure game, is that writing words is a flow, while graphics must be built.
Speaking of – I had a great time building this system. There is a lot of fun space transformation involved in getting the camera’s frame to animate to and from the right place so text is always equally legible, to resizing the frames to fit the text nicely, and subtle, stateful logic when choreographing each sequence of interlocking events (especially when dialogues are interrupting each other and returning, as in the dinner scene). I just want the engineering to serve as a platform for the freeform exploration that’s possible with text and impossible in any other kind of digital game experience.
If I were to try and revise the system for use in a second game, I would build in a more efficient way to track whether there is anything else “of interest” contained within an interaction after one has engaged with it, so players waste less of their time examining information that they’ve already seen. This was a frustration mentioned a couple of times in the feedback.
Others observed that much of the game might as well have been done in Twine, and wondered whether the custom UI was necessary at all. I broadly agree with that observation. I would want to remedy that by building a richer, more autonomous environment, where agents in the world move around on their own, interact with one another in front of you, and can provoke an interaction from you proactively. In other words, I would try to bring in the liveliness of an Inform simulation (if not the wonderful, expressive way one is cooked).
I will also reconsider my use of audio. No one mentioned the audio in the reviews, so I guess it made a mostly neutral impression if it was turned on at all. My own impression is that the spatialized environmental audio (the sound of rain, the soundtrack, the creaking of boards in the burnt out house, the footsteps on different materials, etc) is useful to help conjure an imaginary space, especially in 3D. However, much of the incidental foley felt tacked on, and especially out of place when none of the dialogue was voiced. I would not bother much with these punctuating sound effects if I were to do this again.
The story has its two halves – “love” and “war” – and theoretically they are conjoined by the question of when either of them can be justified. I agree with critics who pointed out that this renders the story diffuse; perhaps it should have been about just one of these things (love), but I was more laissez-faire when writing this because almost everything about the game’s development already had the tentativeness of groping through an unfamiliar and unlit room. Because the topics are now so intertwined, I’ll let this problem lie.
There are some slight spoilers ahead.
One thing I would really like to redress is the question of incest. I’m not at all interested in being edgy or in the cliché’s potential for shock value, so I was chagrined with how it crowded out the other themes. I’m okay with about half of readers thinking that Joe and Zoe are sexually interested in one another, but no more, because I want better representation for the other perspective: that they’re afraid that they are because most everyone else in their lives thinks so.
It’s a mean thing to do, on reflection, but I guess I want to lure the player into worrying that they’re into each other, and then pull out the rug and say “they aren’t, but wasn’t it easy to believe? Wasn’t it easy to read in to? Didn’t all of those loaded sentences actually also have an innocent interpretation that we rejected?” All this, to draw a parallel between ourselves and these imperialists who struggle to see love as anything other than a manifestation of possessive desire.
Instead of finding a place to expostulate this theme, which would be heavy handed, I am looking for places where I can tone down the innuendo, so it feels more ambiguous than it turned out to be. One player said that the moment that made up their mind about the siblings being incestuous was the line from the bedroom scene where Joe asks, “What’s wrong? It’s not like I’m going to kiss you.” It really is on the edge, because even after a string of flirtatious but facetious jokes from a family accustomed to blue humor, like the ribald ribbing Admiral Carnation directs at his son in the previous scene, it really is a terribly calculated thing for someone not interested in his sister to say to her. I need something that could be enough to push Zoe into a panic, which is a direction the player can take her in that moment, but that is still ambiguous enough not to solidify the siblings’ relationship as an incestuous one for the player.
Maybe I’m hyper-focusing on that, though, thinking wishfully that I can change the dominant impression with a single line. There are plenty of other opportunities to dampen the innuendo. I intend to go through the text and touch each of them, to try and rebalance the scales so it’s more possible to interpret either way.
The other leap that the story makes is pitching Joe and Zoe’s confrontation back at their ancestral home as a mortal conflict. For this to work, Joe and Zoe have to be both deeply horrified at the prospect of a seduction and credulous enough about it after reading the love letters to take action. The more guilty they feel about loving each other, the more credence they can give to the other’s having succumbed to desire. Amp this up enough, and I believe they can plausibly pass a threshold where they decide it would be more of a mercy to kill or be killed than continue as lovers. One way or another, it’s a melodramatic turn. I feel like the the melodrama has peers in Lavina’s ridiculous arson plot, Jameson’s subterfuge, and Marion’s medusa-like stalking of the mother, but it would be nice to make such a central shift in the protagonists’ goals a little smoother. I started writing the story with this “dueling snipers with a history” set piece without knowing where it would go, and justifying it obviously involved contortions.
I regret a couple of structural decisions: first, I thought that having a couple of light “puzzles” would give the game some breathing room and help it feel more game-like. These were Zoe finding Joe after missing her shot, and Zoe having to find a way into the kitchens before Joe’s dinner party. My impression is that these were the moments when I lost most players, as they wandered around, got lost, and lost interest. Secondly, making the acquisition of the toy gun at the beginning of the story so influential at the end was a dumb idea to begin with. I just wanted to test the ability of my system to persist game state across chapters, and that damn toy gun kept showing up in the prose until it was plot-critical.
For the post-comp release, I will add more hint-like language (like the blood trail someone suggested) to cut down on the amount of time wasted in these sections. I’ll also add more locations where the gun can be acquired. This doesn’t solve the cruelty of the logic, but since either outcome is, I hope, narratively satisfying, I’ll live with it and let learn.
An enormous thank you to everyone who wrote feedback about “Out of Scope.” I gained a great deal of insight from the experience and had mainly discomfort to offer in return. That, and my considerable gratitude. It is the right day for that, at least here in the United States.