-
Notifications
You must be signed in to change notification settings - Fork 0
Eclipse 2026 Substrate Architecture
(shadow geometry, ionospheric response). The parent owns exactly ONE binding shape Σ_P — the genes-only parent disk shared byte-identically by every invariant class the substrate seals. The two sides of the pair:
- Target descriptor (H1 — geometry ONLY): the quantized obscuration timeline — NASA Besselian element polynomials + station coordinates, computed by the Explanatory-Supplement local-circumstance algorithm, quantized ONCE (obscuration in ppm on a 10-s lattice, contacts in whole seconds UT), stored, fingerprinted. The measured data never touches this side.
- Measured side (the box contents): the station-day's vertical-TEC series, quantized ONCE to milliTECU integers, binned per 300 s (Madrigal's native cadence), with a baseline vector built by the pre-registered rule: element-wise lower median over named quiet control days (max Kp ≤ 4) — an order statistic, never an average, so the arithmetic is integer-exact on every machine.
A width-1 leaf — one UT second t — audits iff:
obscuration(t) ≥ 800000 ppm (the gate)
AND 10 · 10⁶ · (B − M) ≥ 3 · obs(t) · B (the coupling, cross-multiplied Int64)
i.e. the TEC depletion fraction must reach 3/10 of the obscured disk fraction, evaluated only where the geometry says ≥80% of the Sun is covered. Absent bins (−1) never audit; a fake zero cannot exist. The constants live in the sealed Form's decisionPredicate text, so the contentFingerprint pins them — moving them after a seal is a visible falsification, not a tweak.
The obscuration timeline is unimodal: one shadow cone → one contiguous gate window [gateLo, gateHi) per station-day, derived at construction from the timeline alone. regionLinks(patch:) therefore has a genuine wholesale exclusion:
- patch ∩ gate window = ∅ → the whole patch is a proven void — every leaf's C⁴ face is not-whole, its clasp lifts clear of z=0, lk = 0 — ARC frees it in O(1), never split;
- straddling wide patch → subdivide (8-way octant work decomposition);
- width-1 leaf → the audited pair law, then
DeadCatRemoval.isEntangledon the leaf's M⁸ face — the same uniform verdict every substrate invariant class routes through.
Measured behavior: a negative station-day dies in ~1,400–1,900 wholesale voids with zero leaves sealed; a positive seals exactly one leaf — the whole child that threads Σ_P. No number-mixing process exposes such an exclusion from a boundary — the wholesale void is a gift of real three-dimensional geometry: one cone, one contiguous window. This is the substrate's first shadow-projection-class invariant.
pointForEclipseSecond builds one M⁸ point per (geometry, response) cell: S⁴ position = SHA-256 limbs over the pair's own coordinates (station | eclipse | t | measured | baseline) — distinct per surviving leaf; all four C⁴ axes sit at their preserved position iff the pair law holds (c1 1/1 · c2 1/1 · c3 1/1 · c4 256/256), governing axes [c1Reconvergence, c2Boundary] — photoionization reconvergence under the shadow boundary. The face is the substrate's universal one, so the same dead-cat verdict reads every invariant uniformly.
One lane, laneExtentBits: [17] — 2¹⁷ = 131,072 ≥ 86,400 UT seconds of day; leaves ≥ 86,400 are out-of-box dead. The descent is TheoreticalBox → C4Patch octants → leaf, driven by EclipseProjectionRunner.project (stats: leaves sealed, patches voided, max depth, ns).
| Form | Kind | Class | Closure | Terminal | Fingerprint | Stamped |
|---|---|---|---|---|---|---|
| 2024 benchmark | projection | shadow-projection | 7/7 | CALORIE | 89afd825011f060e |
2026-07-16T16:01:47Z |
| 2026 prediction | prediction-2026-08-12 | shadow-projection | PREREGISTERED | CALORIE | c3e6841a206b8058 |
2026-07-16T16:02:00Z |
| 2026 prediction extensions | prediction-extensions-2026-08-12 | shadow-projection | PREREGISTERED | CALORIE | b106d2926b1bf2cf |
2026-07-16T18:54:39Z |
| historical universality | historical-universality | shadow-projection | 16/20 | CURE | cdf43e80f06f10a3 |
2026-08-18T16:42:21Z |
Loaded and fingerprint-checked from ~/.gaiaftcl/franklin/invariants/ at render time.
Three Forms, three scopes, one law:
| Form | Corpus scope | Re-verify re-runs |
|---|---|---|
projection (2024 benchmark) |
2024 eclipse-day + quiet controls ONLY (fp-stable as the corpus grows) | the full crucible over that scope |
prediction-2026-08-12 |
none yet — geometry + constants pinned | byte-exact re-derivation of the pinned text from the elements |
historical-universality |
EVERYTHING — all eclipses + shear negatives | the grouped multi-eclipse crucible |
Every Form carries the Wittgenstein C⁴ tuple {1/1|1/1|1/1|256/256|c1Reconvergence}, class shadow-projection, and a content fingerprint over every semantics-bearing field. Live surfaces (eclipse-watch) re-prove the benchmark Form every pass and refuse on drift (LAW 3).
| Piece | Path |
|---|---|
| Besselian elements (4 eclipses, embedded verbatim) | Sources/HeliosphereIngest/BesselianElements.swift |
| Circumstance calculator + quantized timeline | Sources/HeliosphereIngest/EclipseCircumstanceCalculator.swift |
| Published-circumstance fixtures (trust gate) | Sources/HeliosphereIngest/EclipseGeometryFixtures.swift |
| Madrigal / GFZ / USGS / SWPC clients |
Sources/HeliosphereIngest/MadrigalClient.swift, GeomagIndicesClient.swift
|
| The single float→int crossing | Sources/HeliosphereIngest/IngestQuantizer.swift |
| Ledger store (V252 tables) |
Sources/HeliosphereIngest/HeliosphereStore.swift, Sources/GaiaFTCLCore/HeliosphereSchemaV252.swift
|
| Fossil corpus (multi-eclipse + shear) | Sources/HeliosphereIngest/EclipseFossilCorpus.swift |
| The predicate (RegionInvariant) | Sources/InvariantCompiler/EclipseCollapsePredicate.swift |
| The projection runner | Sources/InvariantCompiler/EclipseProjectionRunner.swift |
| CLI verbs |
Sources/GaiaFTCLCLI/Commands/HelioIngestCommand.swift, EclipseGeometryCommand.swift, EclipseInvariantCommand.swift, EclipseWatchCommand.swift, EclipseWikiRenderCommand.swift
|
| Verify/boot seam |
Sources/GaiaFTCLCLI/Commands/InvariantVerifyCommand.swift (--game eclipse) |
| Live-loop plist (staged) | cells/launchd/com.gaiaftcl.franklin.eclipse-watch.plist |
| Constraint / LAW | How the eclipse invariant satisfies it |
|---|---|
| C1 bond ≤ 4096 | 17-bit single-lane box; descent depth ≤ 17 |
| C3 no FP in sealed proofs | quantize-once discipline; the sealed path is Int64/BigInt only; zero REAL columns in V252 |
| C4 no stubs | every path computes or refuses with the reason; absent data stays absent |
| C5 deterministic verify | order statistics not averages; FNV/SHA fingerprints; fossils jsonl is the canonical artifact (no re-run trig on the verify path) |
| C6 reversibility | append-only ledger rows; Forms replaced only by re-SET, never edited |
| LAW 3 no symptom-masking | falsifiers recorded raw; CURE never renamed; shear false-positives (if any) published |
| LAW 4/9 no classical search | the descent is patch subdivision + wholesale voids; the only "loop" over data is ingest-side quantization of a continuous lattice |
Every data table on these pages sits inside <!-- GAIA:BEGIN … --> markers and is regenerated by gaiaftcl eclipse-wiki-render straight from substrate.sqlite, the fossil corpus, and the sealed Forms — deterministic ordering, idempotent (a second run with an unchanged ledger produces a byte-identical wiki), with --check as the drift gate. The prose is authored; the numbers are the substrate's own.
- The lattice holds
- Impact study — continuum dead
- Death of continuous shear
- Fourier Phantom — Anima FNO vs 11+12+13
- Stellar dynamo kill shot
- QCD: freedom is dilation
- Look in the UI (no visitor data)
- Explore the live courts
- MCP clients (public)
- Example app — entire court
- Math Court user guide
- Math Court on Glama
- Glama connector
- Zero Float · Zero Shear
- UUM-8D vs IUT — WIN
- Readers’ guide
- White paper
- Program index
- Language-games study board
- Known discoveries — family ledger
- Discoveries by user
- Low-friction user flows
- How the IDE hologram works
- Known molecular discoveries
- Study 14 — PDB play
- Study 11 — Ehrhart Volume
- Study 12 — Parallel Repetition
- Study 13 — Connes Rigidity
- Peer-review bundle
- Conjecture alignment
- Study 07 — Milky Way Results
- Study 06 — Explosion Results
- Study 10 — Fermi / Dark Matter
- Study 04 — Tsunami (partial)
- Study 09 — Convective bond
- Study 14 — Protein lattice
- Study 15 — Skala DFT shear
- Study 16 — Disease type
- Study 17 — Chemistry InChIKey
- Study 18 — Material STD
- Study 19 — Go First Dice
- Study 20 — Rife frequency
- Study 21 — Stellar dynamo
- Study 22 — Ground state is a coordinate
- Study 23 — Spin glass, no freezer
- Study 24 — Molecule is consistent or not
- Study 25 — Supremacy is a rounding error