jeanyoon.ch@oi ~ % cd /oi

Welcome to jeanyoon.ch/oi

Jeanyoon Choi (b.1999) is a Computational Artist researching in interconnectivity and complexity through Multi-Device Web Artworks.

You can learn more about my:

↑ ↓ select · enter open · or click / tap
artworks projects experiments research about texts contact

jeanyoon.ch@oi /oi % cat from-euclidean-to-mandelbrotian-making-system-art-natural

From Euclidean to Mandelbrotian - Making System Art 'Natural'

2025-04-15

Reading Geoffrey West and Mandelbrot, I return to the roughness of the natural world. Smooth Euclidean geometry is a poor account of much of what surrounds us. Why, then, should interactive art behave like a neatly manufactured object: deterministic, polished, and limited to a finite set of scenarios?

I want to move from the Euclidean toward the Mandelbrotian, from linearity toward nonlinearity, from flatness toward something uneven. A system could oscillate, persist, and evolve without waiting for a visitor to activate it. Financial markets offer one reference: their variables keep moving even when no single event appears to command them. Weather offers another. The absence of audience input need not mean that the output becomes still.

A conventional interaction often has three phases: a stable condition before the input, a change while the input continues, and another stable condition afterwards. I imagine a different sequence: an already fluctuating system, a period when an interaction modifies that fluctuation, and a new pattern of ongoing movement after the interaction ends. The event changes the conditions of the system rather than simply pushing an object from one fixed state to another.

This helps me distinguish an object from a system. An interactive object may reliably produce the same result from the same action. A system can let an action change not only its visible state but the mechanisms and relationships that generate subsequent states. An automatic system continues to evolve even without an external input.

Adding a noise node in TouchDesigner can make a surface look less artificial, but visual noise alone is not enough. I am interested in uncertainty within the system and within the meaning of an interaction.

There are several possible approaches. One is to imitate randomness directly, which is easy but limited. Another is to simulate many agents that act, trade, or relate to one another in real time. A third is to use a mathematical model of their collective behaviour when simulating every agent is impractical. Brownian motion or a probability distribution might be appropriate in a particular case; arbitrary randomness is not automatically a model of a living system.

The question remains how to construct this kind of ongoing, internally active behaviour. Volatility is central: a work that moves before we approach it, changes through our intervention, and continues beyond us.

↑ ↓ select · enter open · or click / tap
Read original notes back to ../texts ../ back to oi