What is Fractalish?
Fractalish asks whether visible form can preserve process memory strongly enough to support disciplined comparison. It studies roads, cracks, vessels, roots, lightning, surfaces, channels, signals, and other systems as possible records of prior constraint, damage, adaptation, and recovery.
It does not claim that all morphology is readable, and it does not collapse every pattern into one magical fractal explanation.
Accidental longitudinal morphology
Most institutions did not set out to build morphology registries. They simply kept taking pictures while inspecting roads, bridges, forests, bodies, factories, and landscapes. Fractalish treats those secondary-frame records as a potential public evidence layer.
Repeated images beat perfect images. Sequences explain what one frame can only suggest.
Fractal dimension is not enough
Box counting, lacunarity, multifractal descriptors, and boundary metrics can be useful. None is sufficient by itself. Two images can share a fractal-dimension neighborhood while preserving very different process histories.
See the microglia precedent note for a concrete reminder that subtle morphology demands multiple descriptors and acquisition guardrails.
Every claim travels with its trace
Fractalish is not satisfied with aesthetic resemblance. The working standard is evidence packages: image, segmentation, trace, metrics, routing, and provenance. That is how a comparison becomes inspectable rather than impressionistic.
Desiloizing geometry
Road engineers, cardiologists, metallurgists, foresters, and signal researchers often describe related geometry with incompatible vocabularies. Fractalish tries to build the translation layer, not by erasing domain knowledge but by reconnecting it.
Core field guides