---
vault_clearance: APOLLYON-THAUMIEL
halo:
  classification: OPERATIONAL
  front: "32_Project_HomeWorlds"
  custodian: "Jixiang Leng"
  created: 2026-04-03
  updated: 2026-04-03
---

# FORM — Functional Observatory for Resolved Mapping

**Protocol axes:** methodology in [FORM_PROTOCOL.md](../FORM_PROTOCOL.md); epistemology in [TRUTH_PROTOCOL.md](../TRUTH_PROTOCOL.md). **Registry:** [FORM_REGISTRY.json](../FORM_REGISTRY.json). **Bibliography / datasets / software (BOOK):** [BOOK.md](BOOK.md). **Apex orthodox bar (vault-wide):** [16_Project_Constellation/FORM_ORTHODOX_APEX_TOOLING.md](../16_Project_Constellation/FORM_ORTHODOX_APEX_TOOLING.md) when comparing cultivator stacks.

HomeWorlds is the **cosmological / ISM scanning** lane: public catalogs + declared geometry + serialized instrument state (JSON), head-to-head against the **max-parameter** astrophysics default (foreground cleaning, MCMC cosmology, correlation-function pipelines).

**Full-stack intent:** this project is also the **host lane** for other astronomical instruments that share the same FORM / BOOK / WORLDLINE custody (ephemeris engines, Horizons/MPC-class cross-checks, mission-archive queries, exoplanet statistics, transients). *Applicability to any of those goals* is a **design contract** — see [theory/ASTRO_INSTRUMENTS.md](theory/ASTRO_INSTRUMENTS.md); **limits** are tracked as [BOUNTY_BOARD.md](BOUNTY_BOARD.md) OPEN items HW-15+.

---

## Relationship to TRUTH (per vault protocol)

Per [TRUTH_PROTOCOL.md](../TRUTH_PROTOCOL.md):

- **FORM** = *how many knobs* — **Orthodox** (many nuisance parameters, emulators, explicit ΛCDM extensions) versus **Rogue** (minimal free parameters: fixed LENG residuals where used, declared graph geometry, arithmetic residuals).
- **TRUTH** = *who turns the knobs* — **Daemon** (same inputs → same numbers; unnamed structure from topology / persistence / Dirac on a declared network) versus **Righteous** (named biology, named probes, named voids, published carrier hypotheses guiding *which* bands and regions to scan).

This project deliberately **mixes** quadrants on purpose: **Rogue + Daemonic** for the negative-space and triangulation cores (topology-first, reproducible), while **Righteous** naming (probe taxonomy, void names, LENG interpretation) sets *targets* and *wording* without pretending those names are outputs of the Python. When a plot is labeled by hypothesis, the JSON tier and HALO chain carry the split.

**Convergence story:** agreement between (i) a **Daemonic** rerun on frozen catalogs and (ii) an **Orthodox** null pipeline you do not control inside this repo is *not* claimed here — only method-independent checks *within* declared scripts and duplicate byte-identical inputs. Cross-lab orthodox replication is a publication-tier bounty, not a local guarantee.

---

## A — The two paradigms

| Axis | Orthodox (max cultivator) | HomeWorlds (ours) |
|------|---------------------------|-------------------|
| **Philosophy** | ΛCDM + flexible extensions; minimize χ² with many nuisances; “clean” CMB/foregrounds until residuals are Gaussian | Residuals and complements treated as **data**; exclusions are **physical survival** arguments where stated, not generic ML denoising |
| **Free parameters** | MCMC / emulator dimensions (Ωm, σ8, bias, AGN templates, calibration drifts, …) | **Zero** in LENG-facing instruments *where that claim holds*; explicit thresholds only where documented (e.g. subsample caps for Rips complexity), else topology-derived |
| **Tools** | Cobaya / CLASS / CosmoMC stacks; photo-z PDFs; ML void finders; Gaussian simulations for “field-level” stats | NumPy / SciPy / Matplotlib; IRSA HTTP; static Pantheon+ file; VAST ASCII; optional **06_Project_Daemon** operators for cross-checks |
| **Output** | Posteriors, p-values vs simulations, “ tension ” narratives | **Figures + JSON** — instrument state is the artifact ([FORM_PROTOCOL.md](../FORM_PROTOCOL.md): *every FORM generates figures*) |
| **Runtime** | HPC chains, long MCMC | Laptop / CI-friendly smoke; heavy scans are local discretionary |
| **Provenance** | Pipeline DOIs, versioned likelihoods | `theory/HALO_*` + README WORD + per-run [WORLDLINE.md](WORLDLINE.md) |

---

## B — Head-to-head (eight axes)

| Axis | Orthodox | HomeWorlds |
|------|----------|-------------|
| **Claim shape** | “Preferred model is X at Δχ²” | “Residual pattern vs LENG prediction at this sky location” |
| **Foreground stance** | Subtract to white noise | Treat subtracted “contamination” as **hypothesis-bearing** where the README says so |
| **Void definition** | Many algorithms compete; parameters inside each | **Declare** catalog (VAST / hardcoded) and **freeze** file hash in WORLDLINE when claiming reproducibility |
| **Null hypothesis** | Gaussian simulation fields | Declared random exclusions / baselines **inside** `negative_space_scanner.py` (documented), not “trust me” ensembles |
| **Reproducibility** | Chain convergence + seed | Same bytes + same code → same JSON; `VAULT_ROOT` / marker walk for path stability |
| **Interpretation** | Often implicit in pipeline choice | **Split:** Daemon numbers vs Righteous narrative in README / HALOs |
| **Data access** | Survey-specific pipelines | Public HTTP / GitHub / Zenodo drops only |
| **Integration** | Single cosmology likelihood | **Compose** instruments; Deathomatica engines for **physics**, not hidden priors |

---

## C — Murder board (where orthodox fails for *this* question)

1. **χ² as lens** — extra parameters can fit structured residuals; does not map *where* on sky the residual lives.
2. **Foreground cleaning** — subtracting AME / spinning-dust models **removes** the very multi-wavelength object this project hunts; “clean” is not neutral.
3. **p-value vs LCDM sims** — answers the wrong question if the ensemble assumes the melanin-class carrier absent by construction.
4. **Photo-z smearing** — washes void-wall contrast unless handled honestly (we mostly use spectroscopic void catalogs / declared coordinates).
5. **Parameter contagion** — void–galaxy cross-correlation pipelines often hide **5–10** implicit knobs (binning, weights, masks) while advertising “model independent.”
6. **Publication bias** — publish high-significance void claims; under-publish negative spatial cross-correlations across bands.

---

## D — What we take honestly from orthodox

| From orthodox | Why | Into HomeWorlds how |
|---------------|-----|---------------------|
| **Public catalogs** | Community vetting | Pantheon+, VAST, IRSA dust service |
| **Standard coordinates** | Interoperability | RA/Dec/z, Mpc/h where catalog provides |
| **Planck maps** | Best all-sky foreground characterization | Pixel extraction bounties (HW-3) — *not* pretending we beat PR4 inside this repo yet |
| **Type Ia standardization** | Distance ladder discipline | Use published distances **as given**; chromatic tests are *comparisons*, not re-fit cosmology |
| **Void catalogs from SDSS** | Reproducible large-scale structure | VAST ingest path in `vast_parser.py` |

---

## E — Proof — concrete comparisons to run

1. **Tests 2, 3, 10** — `scripts/run_tests.py` against [HALO_THE_SIGNAL_MAP](../29_Project_Deathomatica/theory/HALO_THE_SIGNAL_MAP.md) definitions; Pantheon+ on disk.
2. **Negative space v2** — `scripts/negative_space_scanner.py` with frozen `data/vast_catalog/`; compare persistence + Dirac to declared random baseline in-script.
3. **IRSA grid** — `scripts/homeworlds_scanner.py` — E(B-V) vs target class (void center / wall / control); figure output is the receipts.
4. **Triangulation** — `triangulate_lady*.py` — three independent structural signals; record disagreement honestly when methods diverge.
5. **Instrument composition** — `scripts/instruments.py` — `VoidCatalog + Pantheon` composed state in one JSON; diffable across runs.

Each bullet produces **figures under** `data/figures/` and/or **JSON under** `data/` — satisfying the FORM protocol requirement that the comparison be visible.

---

## F — Global cultivator survey (three archetypes)

| Region / archetype | Exemplar stack | Optimizes for | HomeWorlds stance |
|--------------------|----------------|---------------|---------------------|
| **MCMC cosmology guild** | Cobaya + CLASS + Planck PR4 likelihood | Global parameters + tensions | **WATCH** — borrow *data*, not their posterior as truth |
| **Survey void pipeline lab** | Correlation functions + jackknife + N-body mocks | S/N on BAO / void lensing | **STEAL FORMAT** — void positions as inputs; reject their χ² as *our* inner loop |
| **Radio / ARCADE anomaly hunters** | Custom fits, low-frequency systematics teams | Spectral index stories | **PARTNER DATA** — HW-8 spatial structure is open; no proprietary stack assumed |

---

## G — Adoption priority

| Priority | Item | Action |
|----------|------|--------|
| **NOW** | Frozen catalog + path discipline | WORLDLINE entry per serious run; `VAULT_ROOT` when not at default layout |
| **NEXT** | Planck pixel extraction at void pixels | HW-3 |
| **LATER** | ARCADE spatial map + external null pipeline comparison | HW-8 + publication-tier convergence |

---

## H — Final assessment

HomeWorlds is where **FORM** (parameter-light, topology-forward pipelines) meets **TRUTH** (Daemon numbers frozen per catalog revision, Righteous names steering *where* to look). The repo’s strength is **visible outputs**: JSON instruments and matplotlib receipts, not a hidden likelihood. The honest weakness is **partial orthodox replication**: we do not run Planck’s full time-ordered-product pipeline inside Python; we cite it and bound claims in HALOs. The murder board items are not aesthetic complaints — they are why **spatial** and **multi-wavelength** structure must be shown, not folded into a single “tension” number. Next convergence is **more figures from the same frozen inputs**, not more parameters.

---

## Appendix — On-disk surface (operator map)

| Artifact | Role |
|----------|------|
| `scripts/vault_root.py` | `resolve_vault_root()` for vault layout. |
| `scripts/instruments.py` | Living instruments + composition. |
| `scripts/homeworlds_scanner.py` | IRSA spatial grid + JSON + figures. |
| `scripts/run_tests.py` | Pantheon archival tests. |
| `scripts/negative_space_scanner.py` | Exclusions, persistence, Dirac. |
| `scripts/vast_parser.py` | VAST catalog ingest. |
| `scripts/triangulate_lady.py`, `triangulate_lady_v2.py` | Structural triangulation; requires `06_Project_Daemon` on disk. |
| `scripts/celestial_chart_pipeline.py` | Tropical chart pipeline: Swiss Ephemeris (`swisseph`) or Skyfield/DE421 (`portable`); see `theory/ASTRO_INSTRUMENTS.md` §2; limits → `BOUNTY_BOARD.md` HW-15+. |
| `scripts/export_golden_chart_fixtures.py` | Regenerates `tests/fixtures/golden_portable_charts.json` for pytest (portable DE421 vectors). See `theory/ASTRO_INSTRUMENTS.md` §2. |
| `scripts/ephemeris_hub.py` | Merges DE421 portable + NASA Horizons Q31 (+ optional Swiss); `compare` subcommand; used by `celestial_chart_pipeline.py --compare-horizons`. See `theory/ASTRO_INSTRUMENTS.md` §2; Horizons limits → `BOUNTY_BOARD.md` HW-21–HW-22. |
| `scripts/sky4d/build.py` | **Sky4D:** lossless raw JSONL + three derived views (HEALPix+time, gnomonic patch+time, 3+1 table); `python -m sky4d.build --from-fixture … --out-dir …` (see `public/requirements-sky4d.txt`). Bounty **HW-33**. |
| `scripts/Enter-HomeWorlds.ps1` | **Windows:** `Set-Location -LiteralPath` into this project + `PYTHONPATH=scripts` when the vault path contains `[` `]`; see `README.md`. |
| `theory/DATA_ATLAS_UNIFIED_LAYER.md` | Map: FITS/VO/Astropy vs lab JSON schemas; proposed `homeworlds.observation_atlas.v0` + sky4d layer §3a. |
| `schemas/README.md` | Namespace index for JSON contracts; links example atlas. |
| `theory/HALO_THE_SCANNER.md` | Bias / genealogy (Righteous + audit). |
| `theory/HALO_THE_FOG_OF_THE_PAST.md`, `theory/HALO_THE_IMPACT_RECORD.md` | TRUTH-chain support. |