Super pleased to announce that a custom version of De/Frag (2021) will be on display at ZKM Karlsruhe as part of their new milestone show "The Story That Never Ends" from next Saturday, April 5 until September 20. The exhibition features approx. 100 works from their collection of 12,000 works of the past 70 years. It's a true honor to be part of it!
https://zkm.de/de/2025/04/the-story-that-never-ends
Unlike the original release, which was constrained to only 640x640 pixels, this new custom version runs at 1080x1920 resolution on a large 75" screen, and will continuously evolve over the course of the next 5½ months, i.e. for the entire duration of the exhibition.
De/Frag is a homage and exploration of the classic flood fill algorithm, combined with semi-randomly generated bit/fill patterns, multiple iteration & sorting strategies, creating an endlessly changing tapestry of complex pixel textures and tectonics, whilst revealing the inner workings of the intentionally slowed down core algorithm. It also is an exercise in algorithmic purity, attempting to devise a minimalist generative system producing maximalist variation (relative to its setup).
The downscaled 30 second time lapse shows snapshots of a 20 minute time window, with one frame taken every 5 seconds... In reality, all transitions are continuous and mesmerizing to watch unfold in unexpected way (something the time lapse cannot do justice).
For anyone interested: Here's an ambiguous set of instructions (in the spirit of Sol leWitt or https://conditionaldesign.org) for (re)creating a simplified version of the infinitely evolving De/Frag piece (see beginning of this thread).
Preparations:
- You may use any suitable medium.
- Create a pixel-based canvas of arbitrary dimensions.
- Create a palette of random choice of 4 colors, plus black and white.
Repeated steps:
- Choose two random colors from the palette, giving black or white priority.
- Using the chosen colors, create a random pixel pattern in one of four sizes (e.g. using percentages of the longest side of the canvas as pattern size, e.g. 1%, 2%, 5%, 10%, or 4x4, 8x8, 16x16 pixels...)
- Choose a random pixel position on the canvas and a random color tolerance.
- Perform the flood fill algorithm[1] from the chosen start position, selecting all connected pixels within the chosen tolerance.
- Apply the new pattern (created in step 4) to the pixels selected in step 6, in an order and speed of your choosing.
- Repeat from step 3
[1] The Flood Fill algorithm is described here: https://en.wikipedia.org/wiki/Flood_fill
