Not a gallery of results.
Working benches.

These are the models on which we learn to ask precise questions about connections. Each has a version, observables, limits of applicability, and a history of unsuccessful moves.

A card appears here not when the work is “finished,” but when it can be described honestly enough for another person to understand what is being tested.

concept sandbox working model reproducible result archive

M–01

Geometry and structuresworking model

VSEPR: shape as a result of constraints

An attempt to turn familiar geometric intuition into a computable object. A configuration emerges from competing constraints rather than being supplied as a ready-made picture.

Main question: Which shape classes follow robustly from the chosen relations and constraints?
Observable: Geometric class, symmetry, and classification stability.
Honest limitation: Matching a shape alone does not explain the physical nature of a relation.

Research Square preprint · version 1 ↗External page on Gist.Science ↗

M–02

Dynamicssandbox

Three bodies: what remains observable

A sandbox for separating an attractive trajectory from a robust signature of a regime. We ask not only what happened, but whether the classification survives a change of representation.

Main question: Which dynamical properties survive admissible coarse-graining?
Observable: Motion regimes, transitions, and label stability.
Honest limitation: The sandbox studies a language of comparison; it does not promise a general analytic solution.
Public materials: The working card is being prepared for publication.

M–03

Morphology and graphsworking model

Structural sandbox

A move from visual similarity to graphs, controlled quotients, and transition corpora. This is where it becomes especially clear why some comparisons must return INVALID.

Main question: What is essential to a structure, and what belongs only to its representation?
Observable: Graph class, admissible transitions, and preserved distinctions.
Honest limitation: The result depends on an explicitly chosen language of observation.

Why this required CQT →

M–A1

Early numerical layerarchive / review

Rotation curves

An early numerical branch of the project, kept visible as part of its history. It is useful to see not only today’s method but also the problems for which that method was not yet sufficient.

Why it remains: To keep the development of criteria and model language inspectable.
Current status: Not a central branch of the new programme; requires reassessment.

Model description and figures →Source code and numerical summary · ZIP, 2.5 KB ↓

Catalogue rule

We will not label every branch a “model” as though all were equally mature. A status must answer one simple question: what can another person already verify here?