I was playing the game Transistor and was inspired by some of the animations in it, so I decided to try to replicate this in Javascript. The component is a massive mess that I will clean up one day but for now its cool to play with.
Also 0 optimisation has gone into this so far.
A deployed version of the storybook is available here: https://aiden-ziegelaar.github.io/CircuitHero
To get started quick:
yarn
yarn storybook
Whilst it was fun to do, the maths is pretty simple at its core.
For each pip generated there a number of actions it can take on any tick, it can continue moving in the same direction (i.e. no action), it can bifurcate into two pips, it can change direction, or it can die.
The likelihood of any of these functions not decaying in a given tick is given by the halflife formula:
or to convert this to a binary value of if the event occurred in that tick we can check if Math.random()
returns a value greater than this number using the following javascript:
function probabilityFromHalflifeAndElapsed(halflife: number, elapsed: number): boolean {
return Math.random() > (Math.pow(0.5,(elapsed/halflife)))
}
We want to emulate a circuit, hence to make a design reminiscent of a circuit board we constrain the ability of pips movement to the cardinal and intercardinal directions only. This allows us to easily precompute our sine and cosine multipliers:
const directions = Array.from(Array(8).keys())
const sinPrecalc = directions.map((input) => Math.sin(input * Math.PI/4))
const cosPrecalc = directions.map((input) => Math.cos(input * Math.PI/4))
In this fashion we can consider the direction to be represented, not by a degree value, but by the index of the sine and cosine arrays, representing our 8 constrained directions.
For the pips themselves we can specify a very basic model:
interface ILocation {
x: number;
y: number;
}
interface IPip {
direction: 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7;
location: ILocation;
path: IPath;
}
interface IPath {
locations: ILocation[];
}
We can see our constrained directions indicated as the index of our sine and cosine array. To trim the tail all we do is add the distances of the locations in the path and work out the distance that the final element should be, either modifying the final element, or removing the last element and modifying the new last element in the case that a tail goes around a corner.
The trails have a lifetime and their length is determined by the velocity of each particle and the trail lifetime. For a pip that has been stopped, we simply add a timestamp to the pip of when it was stopped, then trim the tail by the difference in the stopped time and the elapsed time. Once the length of the tail reaches zero we remove the pip from the canvas.