This project is my first one in a series of algorithmic studies. Inspired by the work of Casey Reas and his "Process Compendium", it starts with a very simple text description of an algorithmic process. In this case, I came up this description:

"Place N circles randomly on the canvas. Select a point on each of the circles, and rotate it around its circle. At each step, draw a random line that connects all the points"

Each step is drawn with very low opacity, and is repeated a large number of times (these pieces draw around 6 million lines). Eventually, complex structures, shapes and patterns appear, that were impossible to predict from the text description alone: this is an emergent system.

The text description voluntarily leaves things open to interpretation: how many points ? which colors to use ? which rules for the drawing opacity and line weight ? Do all points move at the same speed ? Can they accelerate ? Are the circles allowed to extend past the canvas boundaries ?

This makes the process capable of interesting variations, as well as reusable: another artist implementation of the same description would most likely give very different artworks.

Despite the apparent simplicity of the description, this collection ended up being a real challenge. Drawing millions of lines with low opacity turned out to be impossible without a GPU. WebGL is required.

This token uses the regl.js library. The outputs are rendered at 2400x2400 and should be viewed at that size. You can generate high resolution outputs, up to 8000px, by adding &size=XXXX to the URL.