next »

[dkilmer]

Reproduction Man is releasing tomorrow, so I am ready to move on to the next thing. I'm using "Trifle" as a working title for now.

I have been thinking about this game on and off for about a year. I was wanting to do a bigger project (and I think it will be), but I realized that there's a huge benefit to making it a one-person thing. Plus I don't mind it taking several years to do.

Like a lot of game developers, the game that I want to make is the game that I really want to play but doesn't exist yet (although if it does and I just haven't heard about it, I *definitely* want to know about it). Also, like a lot of game developers, my ideas have been influenced by the games that have affected me most deeply. For Trifle, the influences are Kingdom, Loren Schmidt's recent work on a game called "Goodnight Traffic City", William Chyr's dev streams for Manifold Garden, Toki Tori 2, The Witness, Tetrobot and Co., Infinifactory, and Lemmings. Yeah, that's a weird mix. There are a lot of puzzle games in there, but Kingdom has the most influence, because it was the first "strategy" game that I found myself unable to stop playing or thinking about.

The epiphany for me was a way that I think a resource/strategy game can also be a gardening game and a puzzle game without being a bunch of elements of each slapped together. I'll develop that idea more later, but for now, here is a big list.

What the Game Should Have

• A world that is 2d. You are interacting with the world you see, not a simulation of dimensions that you can’t experience directly.
• A world that is visually interesting, novel and kinetic.
• A world that is an interconnected physical system that operates causally.
• A character that you control who moves around the world. You are not a disembodied presence — you are a thing within the world, subject to all the constraints that everything else is subject to.
• A world that physically changes over time as a result of your actions.
• A bounded world. You are tending the world, shaping it and learning it, not trying to escape it. It needs to be expansive, but not so expansive that you can’t come to know all of it.
• Cycles of growth and decay. Maybe even seasons. Definitely weather.
• Cycles of day and night. The time of day should be important to what is happening.
• A world with areas that are blocked off in the beginning until you acquire the resources and knowledge to open them up.
• No words.
• The ability to create things that weren’t directly designed into the game.
• Gameplay that involves puzzle-solving and situational ingenuity rather than just strategy and resource gathering. But also strategy and resource gathering.
• A sense of relating personally to other characters in the game. A sense that at least some of the other characters are your friends or enemies.
• A world that invites you to understand it. Nothing deliberately obfuscatory or needlessly cryptic.
• A sense of risk. The sense of making a leap without complete knowledge of the consequences, but with trepidation about the outcome.
• A story — the real kind that emerges from the things that happen in the world, but that is constrained by the basic (designed) premise of the world.
• A central mystery relating to the nature of the world and your purpose within it.
• The ability to enhance your own abilities.

I think the hardest part will be the absence of language. How can you get a sense of another character without verbal communication? How can you get a sense of story without at least starting to be told it? For almost everything above, I could name a game that does it. Not that one.

[dkilmer]

Prototyping what the Trifles might look like:

[dkilmer]

Spent some time trying to prototype Trifles. I know that I want them to be "flauna" -- basically, creatures that have limbs and move around but that are "hatched" from seeds. The difficult constraint is that they are going to be very small in game size.

I thought at first that I wanted something gastropod-like, and tried this:



This did not shrink well to a small size. So I tried something more like a humanoid, but with less detail:



Didn't really shrink well either. Plus they don't meet a major criterion -- expressiveness. I want the player to be able to read the Trifles' reactions and hopefully be endeared to them. To make that work, I realized that these creatures need to be pretty plump looking and pretty much all face. So I started working in Aseprite instead of Procreate, and I came up with this:



I am pretty happy with it. One detail that's not yet present is that the seeds (which constitute their bodies) will have a kinetic texture (possibly random, possibly based on some kind of CA). I think I can make it work. When they "hatch", their limbs and eyes will sort of pop out of the seed. And they will retract back into the seed when they sleep or are really frightened.

[dkilmer]

For the last week or so, I've been continuing groundwork. Mainly:
1. Working with Unity to see if it will be bearable or make anything easier.
2. Testing out ways of growing plants.

Unity turns out to be really useful, for prototyping at least. It's not hard to dynamically create and modify meshes, and it's really fast to try things and see if they're viable.

Trees were a conundrum. I want the trees to be varied and genetically determined. An obvious solution is L-Systems, but there are two aspects to L-Systems that make them sub-optimal in my situation:

1. It is really hard to generate rules that have predictable behavior. In a game where the genes will be semi-stochastic, this is probably a really bad thing. Imagine accidentally generating huge or geometrically nonsensical trees.
2. The method of generation is recursive *replacement*, which means it's really difficult to animate the growing process itself. Ideally, I wanted something where a part, once created, would continue to exist and could be animated.

Then I found this lovely paper by Runions, Lane and Prusinkiewicz, which describes a method of generating trees using a "space colonization" technique. A huge benefit of the algorithm is that you define the tree's bounds up front. Another benefit is that there are a lot of parameters (growing distance, attraction distance, pruning distance, shape, branch weight, etc.) that can be encoded in genes within known-to-be-reasonable ranges. After some messing around, I got this:



I'm not happy with the way it grows. What I want is for branches to start sprouting before the parent limb is finished growing. I've got some ideas about how to do that, which I'll post about in detail if I get them working.

[TonyManfredonia]

Wow! I love the latest character.

All three of the character prototypes were great, but I think the bit more "ambiguous" seed-like creature will work well.

The concept sounds fantastic; I love how your personal passion is driving it. Staying posted on this, for sure!

[dkilmer]

Thanks for the encouragement!

[dkilmer]

And here's that second try at tree-growing. This is going to be more difficult than I thought. I think I'll leave it for a while, though, and let it simmer.

[TonyManfredonia]

Right, I can see how that can be tough. It seems like it's probably a fine-balance between the growth of the trunk vs. the branches.

Hmm, maybe let more branches extend before the trunk/base gets larger?

[Thaumaturge]

I don't know how feasible this is (or how realistic, for that matter), but one thought that occurs to me is to have branches only start growing once their parent branch has reached the point at which they're intended to start growing, and likely only once the parent branch has passed a certain base width (perhaps twice the target width of the new branch, or something similar).