The CAELIX Project

CAELIX is a research and simulation framework for discrete lattice field experiments. Its central question is whether continuum-like behaviour can be constructed from local, finite, balanced-ternary substrate rules rather than assumed as the starting language of the model.

The public site is the visible surface of that work: a browser-based experiment suite, a growing document library and a record of the modelling choices that have survived enough testing to be worth writing down. Behind it sits the heavier development track: native runtime work, integer field transport, balanced-ternary microstate seeding, motif interaction logic and GPU-aware simulation tooling.

Experiment Families

The browser suite now spans microstate seeding, substrate settling, propagation fields, interference, confinement, lensing, light-clock behaviour, time reversal, oscillator and breather stability, field-mediated N-body motion, fan-out geometry, voxel gearing and black-hole analogue tests. Some experiments remain deliberately simple. Others are stress tests for whether the architecture keeps its footing when geometry, burden, closure and propagation interact.

The newer experiments are no longer just demonstrations of field effects. They are probes of the CAELIX design itself: whether balanced-ternary seeding produces useful structure, whether SE and PE layers should remain separate, whether integer Laplacian transport can replace floating-point carrier scaffolds, whether thin-slab 3D behaviour differs from flat 2D behaviour, and whether burdened rotating defects can produce measurable closure rather than decorative visuals.

Research and Documentation

The document stack has expanded from site support into a public research trail. It includes the white-paper sequence on existence, balanced ternary, constants, emergent fields, Standard Model stencil structure, substrate audits, bounded integer field transport and voxel gearing. Working notes cover scale boundaries, hierarchy, non-locality, motif interaction, particle mapping, universe seeding, dimensional geometry and black-hole geometry.

Those documents are not ornamental. They are the audit layer for the project. They separate settled implementation, working doctrine, speculative structure and open problems, so the code and the claims do not drift apart. When the experiments change the argument, the documents are updated rather than left as fossilised sales copy.

Method and Design Logic

CAELIX treats simulation as an experimental instrument. The lattice is not a visual grid placed under familiar physics. It is the object being tested: a finite signed substrate whose local rules, registers, stencils, boundaries and propagation operators can be inspected directly.

The working architecture now distinguishes the Substrate Engine from the Propagation Engine. SE handles exact local legality, settling, collapse, motif extraction and contact truth. PE handles higher-order propagation, burden fields, annihilation release, transport, and the question of when exact computation must be paid for again. That split has become one of the main organising principles of the project.

A second organising principle is substrate honesty. Floating-point carriers and browser visualisations remain useful, but they are treated as scaffolds. The current direction is toward bounded integer balanced-ternary registers, explicit energy ledgers, fixed-scale rendering, deterministic seeding and native validation where the browser is too small a box for the experiment being asked of it.

Project Scope

CAELIX now spans a public web suite, a white-paper corpus, working research notes, JavaScript prototypes and a native Rust/Metal runtime track. The web layer remains phone-first and readable, but the deeper simulation work is moving toward a stronger local engine capable of larger fields, better profiling, typed state, GPU/CPU comparisons and eventually more serious 3D validation.

The near-term direction is to keep the public site coherent while tightening the research spine beneath it: cleaner document pages, better experiment provenance, stronger integer transport, clearer SE/PE contracts, native back-end validation and a less forgiving distinction between what has been demonstrated, what is being tested and what is still only a structural clue.