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 system-art-progressive-storytelling-vs-state-based-approach

System Art: Progressive Storytelling vs. State-Based Approach

2025-02-17

When designing interactive and multi-device artworks, we often begin with moving images because they provide a familiar reference for media art. That reference can quietly impose a linear structure: arrive, interact, advance to the next scene.

But an interactive system need not work that way. If every action merely unlocks another predetermined stage, the participant remains a puppet in a storyboard. They may alter visual effects without affecting the dynamics that organise the work.

I want a reciprocal relation. The participant changes the system; the changed system changes what the participant encounters and does next. Actions can affect important relationships in both the short and long term. Narrative emerges from that exchange rather than simply being delivered through it.

This is easier to describe than to build. Complexity may prevent a visitor from beginning. The system itself is difficult to design, and excessive unpredictability can create both conceptual and engineering problems.

I distinguish three approaches. Linear storytelling follows one predefined sequence; interaction chiefly advances it. Progressive storytelling allows branches or parallel states, but the larger structure remains predetermined. A state-based approach treats the work as a network whose states and relationships can themselves change through interaction, potentially including the generation of new states.

SoTA is an example of progressive storytelling. It has three broad conditions—Idle, Frontend and Backend—related to whether a participant is absent, active or inactive. Within the Frontend, the participant navigates 118 neural-network architectures. This produces a layered and somewhat nonlinear structure, but the participant still selects among existing states rather than changing their relationships. It can also be understood as a repeating Idle–Frontend–Backend cycle.

A more fully state-based work would need a different analysis, perhaps using graphs. I would approach that gradually rather than replace a workable progressive structure with unmanageable complexity.

Linear and progressive narratives tend towards identifiable beginnings and endings and often suit one participant. A state-based work can have less definite boundaries and accommodate several people joining an ongoing situation. Controlling a system is not sufficient to distinguish them; the stronger proposition is becoming part of it.

I have not encountered a fully realised example of the state-based ideal I am describing. Some multichannel interactive works suggest aspects of it, but simple cause and effect or the use of live data alone does not fulfil the ambition. It remains an attractive direction for genuinely reciprocal system art.

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