Microstate Seeding Lab

A CAELIX experiment in lawful balanced-ternary seeding, substrate settling and SE/PE runtime handoff

Run Experiment

Open the live browser experiment as a single-run view.

What Is It?

This experiment fully writes a square lattice container from a lawful balanced-ternary constant source before any later runtime behaviour is considered. It is a seeding lab first and a runtime demonstration second.

The default grid is 243 × 243, matching the project’s radix-3 scaling logic. The seeded values are not random speckles. They are trits drawn from a constant tape, currently imported from precomputed balanced-ternary payloads for constants such as √2, √3, π, e, φ and related kernels.

What It Tests

The experiment asks whether different lawful write orders produce measurably different load structure once the same balanced-ternary tape is written into the same finite container. The mapping order is therefore part of the test, not a cosmetic traversal choice.

It also tests the transition from a seeded substrate into runtime behaviour. After the container is filled, the experiment can pass through substrate settling, Substrate Engine evaluation and Propagation Engine transport. That makes it a compact public view of the wider SE/PE architecture.

How It Works

A balanced-ternary tape is decoded into {-1, 0, +1} values and written cell by cell according to a selected mapping order. Available traversals include spiral, row-major, row-snake, diagonal sweeps, diamond shells, 3×3 block recursion, ternary Morton order, Peano-style paths and ring-based clockwise orders.

Once written, the field is measured for occupancy, radial load and anisotropy. The benchmark path can compare axis/diagonal balance, horizontal/vertical balance and angular distribution across grid sizes and mappings.

After seeding, the same substrate may be handed to the runtime layers. SE handles local legality and settling. PE then treats the result as a source for integer balanced-ternary wave transport, using bounded φ and velocity-like registers rather than an unbounded floating-point field.

What Is Not Hard-Coded

No random fill is used in the default path.
No post-hoc smoothing is applied to make the seed look isotropic.
No preferred mapping is declared by visual taste alone.
No runtime behaviour is considered before the seed has fully written the container.

The experiment separates the seed, the write order, the measured load structure and the later runtime layers. That separation is the point.

Why It Matters

CAELIX treats the initial substrate as a structural object, not a random background. If a finite lattice is filled from a lawful balanced-ternary tape, then the order of writing and the geometry of the container can alter the measured burden and anisotropy.

This experiment makes that problem visible. It gives the project a controlled way to compare constant sources, grid sizes and mapping orders before asking the more expensive question of how those seeded structures settle, collide, transmit burden or produce propagating fields.