WP15: proposal package — README first screen, docs/ tree with R16, adoption artifacts

.github/workflows/ci.yml

@@ -54,11 +54,18 @@ jobs:
       - name: Fill README.md marker blocks (tools/readme/fill_readme_numbers.py)
         run: python3 tools/readme/fill_readme_numbers.py .
 
-      - name: Reproducibility gate (committed generated+evidence+README == regenerated, SHA-stable)
+      # WP15 (R16): fills every DOCS-*-START/END marker block across
+      # docs/*.md from the committed CSVs they cite -- the docs/ sibling of
+      # the README step above, same "separate step before the reproducibility
+      # gate" convention (docs/ is not under generated/ or evidence/ either).
+      - name: Fill docs/ marker blocks (tools/docs/fill_docs_numbers.py)
+        run: python3 tools/docs/fill_docs_numbers.py .
+
+      - name: Reproducibility gate (committed generated+evidence+README+docs == regenerated, SHA-stable)
         run: |
-          git diff --exit-code -- generated/ evidence/ README.md \
-            || (echo "::error::generated/, evidence/ or README.md drifted from committed bytes (R6)"; \
-                git --no-pager diff --stat -- generated/ evidence/ README.md; exit 1)
+          git diff --exit-code -- generated/ evidence/ README.md docs/ \
+            || (echo "::error::generated/, evidence/, README.md or docs/ drifted from committed bytes (R6)"; \
+                git --no-pager diff --stat -- generated/ evidence/ README.md docs/; exit 1)
 
       # if: always() -- upload the regenerated bytes even when the gate fails,
       # so an R15 re-baselining PR (a legitimate, documented pin change) can

CMakeLists.txt

@@ -196,6 +196,14 @@ if(PYTHON3_EXECUTABLE)
     add_test(NAME readme_numbers COMMAND ${PYTHON3_EXECUTABLE}
              ${CMAKE_CURRENT_SOURCE_DIR}/tools/readme/test_readme_numbers.py
              ${CMAKE_CURRENT_SOURCE_DIR})
+    # WP15 docs/ marker-block filler (R16 mechanism): regenerates every
+    # DOCS-*-START/END marker block across docs/*.md from the committed CSVs
+    # they cite (the docs/ sibling of readme_numbers above) and checks
+    # --check mode passes against the committed files, catches a corrupted
+    # number, and is deterministic/idempotent.
+    add_test(NAME docs_numbers COMMAND ${PYTHON3_EXECUTABLE}
+             ${CMAKE_CURRENT_SOURCE_DIR}/tools/docs/test_docs_numbers.py
+             ${CMAKE_CURRENT_SOURCE_DIR})
 else()
-    message(WARNING "python3 not found; WP7a viz, WP8 compliance, WP7 evidence, WP10c forensics and WP14 prioritization tests not registered")
+    message(WARNING "python3 not found; WP7a viz, WP8 compliance, WP7 evidence, WP10c forensics, WP14 prioritization and WP15 docs_numbers tests not registered")
 endif()

docs/concept.md

@@ -0,0 +1,168 @@
+# Concept
+
+> This is a concept dossier for a research/concept-study audience — **not a
+> mission proposal**. Nothing here is legal advice, a flight-readiness claim,
+> or a commitment on behalf of any agency. TRL 4, GNC software element only
+> (see [roadmap.md](roadmap.md)).
+
+## What ADSC is
+
+ADSC is an open-source, reproducible **evidence package** for installer-type
+active debris removal, framed as **Kessler-precursor removal**: instead of
+towing large debris or chasing fragments, a small servicer installs a passive
+deorbit kit on the massive derelicts that would otherwise become the next
+fragment clouds, and departs. The audience is mission designers at the
+agencies and contractors of major launching states. The product is
+credibility; the success criterion is that a skeptical engineer can clone the
+repository and regenerate every claimed number
+(`bash tools/regenerate_all.sh build`, byte-identical across three independent
+platforms — see [technical_architecture.md](technical_architecture.md)).
+
+Canonical summary blurb, used verbatim (spec §1):
+
+> ADSC is not a debris collector. It is a low-cost concept for
+> Kessler-precursor removal: a small servicer installs a passive deorbit kit
+> on high-risk massive derelicts and departs, preventing future fragment
+> clouds before they form — maximizing cleanup value per unit cost with a
+> passively-safe approach design.
+
+The owner's one-line governing value for v5: maximum orbital-environment
+cleanup value per unit cost, with approach safety treated as a design
+requirement rather than a statistic. That approach safety is a
+**passively-safe approach design with clearance-verified aborts (keep-out
+violations 0/500 per catalog [L0, ds-v1, Wilson ≤ 0.0076])**, re-verified at L1
+(+J2) and L2 (+drag) — never "strict approach safety" or "safe satellite";
+the full campaign statistics and re-verification tables are
+[safety.md](safety.md)'s to own, not repeated here.
+
+"Minimum cost" as an absolute claim is banned throughout this package. The
+defensible claim is **cost-effectiveness — cost per removal and cost per unit
+risk reduction** (see [cost_model.md](cost_model.md)).
+
+## Kessler-precursor removal — the cascade framing
+
+The collisional cascade's fuel is the population of massive intact derelicts
+in congested orbital bands; fragments are the symptom. One intact-intact
+collision produces thousands of trackable fragments — the 2009 Iridium 33 /
+Cosmos 2251 collision alone produced 598 + 1603 = 2201 fragments cataloged by
+the U.S. Space Surveillance Network as of January 2013 (Liou, J.-C., "An
+Analysis of the FY-1C, Iridium 33, and Cosmos 2251 Fragments", NASA Orbital
+Debris Program Office, NTRS 20150003820, 2014). Removing, or equipping for
+removal, the objects that would become the next fragment clouds attacks the
+source term rather than the symptom — that is what **Kessler-precursor
+removal** names.
+
+Anchors in the open literature:
+
+- Kessler, D. J., Cour-Palais, B. G., "Collision Frequency of Artificial
+  Satellites: The Creation of a Debris Belt", *Journal of Geophysical
+  Research* 83(A6), 1978 — the cascade mechanism itself.
+- Liou, J.-C., Johnson, N. L., "Risks in Space from Orbiting Debris",
+  *Science* 311(5759), 2006 — LEO population growth even without new launches
+  (the instability argument).
+- Liou, J.-C., Johnson, N. L., Hill, N. M., "Controlling the growth of future
+  LEO debris populations with active debris removal", *Acta Astronautica*
+  66(5-6), 2010 — the classical few-removals-per-year
+  environment-stabilization-class result (~5 objects/yr, jointly with ~90%
+  post-mission-disposal compliance — removal alone does not carry the result)
+  that sizes the campaign cadence ADSC targets.
+- McKnight, D., et al., "Identifying the 50 statistically-most-concerning
+  derelict objects in LEO", *Acta Astronautica* 181, 282-291, 2021,
+  doi:10.1016/j.actaastro.2021.01.021 — a consensus ranking (11 expert teams)
+  dominated at the top by SL-16 / Zenit-2 second stages; average mass of the
+  top-50 ≈ 5,295 kg; removing the top-50 would roughly halve LEO
+  collision-risk (top-10 ≈ 30% reduction). This is the citation basis for
+  ADSC's reference target classes (D2).
+- ESA Space Debris Office, "ESA's Annual Space Environment Report",
+  GEN-DB-LOG-00288-OPS-SD, Issue 10.0, 1 May 2026, §7.1 — the
+  environmental-index peak near ~800-900 km / 70-80° inclination matches
+  ADSC's catalog_A reference class (~840 km / ~71°), under a 90%
+  post-mission-disposal-success assumption.
+- Population figures (≥1 cm ~1.2 M objects) are MASTER-8-class model values
+  (MASTER-8, released March 2019, reference epoch November 2016; Horstmann,
+  A., Hesselbach, S., Wiedemann, C., "Enhancement of S/C Fragmentation and
+  Environment Evolution Models", Final Report, ESA contract
+  4000115973/15/D/SR, 2020), carried as PLACEHOLDER inputs in the T6 flux
+  table (`generated/t6_flux_sweep.md`).
+
+These are bibliographic/regulatory citations (years, DOIs, report editions),
+not ADSC simulation results, and are cited here rather than machine-included
+from a committed CSV.
+
+## Installer, not tug — the quantitative argument (D1)
+
+Deorbit Δv scales with the mass it must decelerate. A tug must decelerate the
+full ~9,000 kg SL-16-class stage — roughly 304× the entire ADSC servicer's
+own contact mass (29.6 kg dry+kit; see evidence pack §2) — and a realistic tug
+bus would be heavier still. The installer instead transfers a 2.4 kg passive
+kit and departs, so the propellant cost of the deorbit itself is carried by
+drag (sail) or electrodynamic drag (tether), not by the servicer. One mission
+services several targets in one plane (batch amortization —
+[cost_model.md](cost_model.md) owns the amortization-curve numbers); the
+capture interface is a geometry-keyed clamp on features present on every
+upper stage by construction (nozzle throat / adapter ring, D4), not a generic
+manipulator. This mass-ratio argument is the placeholder-independent core of
+the package: it survives every open parameter, because it is a scaling
+argument, not a point estimate.
+
+The GNC element itself is textbook-level by design (D11): Clohessy-Wiltshire
+relative motion, quaternion sliding-mode tracking, multiplicative EKF — open
+literature methods applied to class-level targets, with no live ephemerides
+and no target-specific operational products
+([gnc.md](gnc.md) owns the mechanism detail).
+
+## What an agency already has — the counterfactual
+
+An agency evaluating this package already has, or can buy, several adjacent
+things. None of them is the installer trade this package makes:
+
+- **A funded capture/tug-class demonstration.** ESA's ClearSpace-1 mission —
+  contracted at **~€86M** (November 2020; ~€100M total project cost), now
+  targeting the defunct PROBA-1 satellite (retargeted from the Vespa adapter
+  in April 2024) — is presently expected to launch **~2028**. This is a
+  single-target capture-and-deorbit tug demonstration, not a multi-target
+  batch-amortized installer, and it does not by itself answer the
+  installer-vs-tug mass-ratio question this package argues quantitatively.
+- **A funded national rocket-body-removal demonstration.** JAXA's Commercial
+  Removal of Debris Demonstration Phase II (CRD2) — Astroscale Japan's
+  ADRAS-J2, contracted at **~¥13.2B** (2024) — targets an unprepared Japanese
+  upper stage, with the demonstration presently planned toward **FY2027–28**.
+  Like ClearSpace-1, this is a single-target removal demonstration by a funded
+  agency program; it validates that agencies will pay for large-object
+  removal, not that an installer architecture is the cost-effective way to do
+  it at scale.
+- **Published passive-kit physics, without an installer-architecture trade.**
+  The bare electrodynamic tether (EDT) literature is mature and citable:
+  Sanmartín, Martínez-Sánchez & Ahedo's 1993 OML bare-tether foundation paper
+  (*J. Propulsion and Power* 9(3), 353-360, doi:10.2514/3.23629), and the EU
+  FP7 BETs project (GA 262972, coordinated by Prof. Sanmartín, UPM,
+  2010-2014), which demonstrates order-of-magnitude tether-mass-vs-deorbit-time
+  results (e.g. a 10 kg tether deorbiting a 1000 kg spacecraft from 1000 km in
+  about a month). Drag-sail deorbit is similarly well published. What this
+  literature does not provide is an open, honesty-audited trade of the
+  *installer architecture itself* — batch amortization across several
+  targets per mission, the geometry-keyed capture interface, and a
+  campaign-level Monte-Carlo safety case with clearance-verified aborts.
+
+**The niche this package fills**: an open, reproducible, honesty-audited
+*installer-architecture trade* — not a new capture demonstration and not a
+new tether-physics result — sitting between the funded single-target
+tug/removal programs above and the published component-level kit physics,
+making the batch-amortization and installer-vs-tug argument in a form a
+skeptical engineer can regenerate byte-for-byte.
+
+## Scope guardrails
+
+- **Operator-legality framing (D9).** The package assumes the operator is, or
+  is contracted/consented by, the launching state of the target. Nothing in
+  this repository assumes or enables unconsented approach to another state's
+  object.
+- **Dual-use guardrail (D11).** Methods stay at open-literature level (CW,
+  sliding-mode control, EKF are textbook material). Targets remain
+  class-level parameters: no mission-planning products against specific
+  catalog IDs, and no ingestion of live ephemerides/TLEs.
+- **Language discipline (spec §1, binding across README/docs/evidence pack).**
+  The anchor phrase **"Kessler-precursor removal"** appears in every summary
+  text, including this one. Adoption-facing deliverables — README, docs/,
+  pitch docs — stay in English (R11); a Japanese equivalent of the summary
+  blurb may accompany it only in owner-facing material, not here.