Interstellar Archive

Thesis

This repository treats long-horizon mission reasoning as an evidence-governed engineering problem, not a narrative thought experiment. The public contract is deterministic computation plus explicit uncertainty, versioning, and governance enforcement. Every non-trivial claim should be reproducible from tracked inputs and scripts, or marked as speculative with visible boundaries. The goal is not to prove optimistic outcomes, but to make assumptions auditable and model drift impossible to hide.

Research Signals

Abstract

This repository publishes a deterministic, schema-driven simulation and evidence workflow for long-horizon encounter analysis. It is structured as a reproducible research artifact and publication candidate. Claims are tied to assumptions, models, artifacts, and sources through reproducible checks. Simulation outputs are guarded by a golden checksum and CI enforcement to prevent silent drift. Governance documents define decision traceability, reproducibility requirements, and explicit non-goals.

Start Here

1. Read the thesis and trust model in this README.
2. Run the fast public checks:

   python3 scripts/ci/required_paths_validate.py --strict
   python3 scripts/ci/validate_research_signals.py --strict --no-require-tag
   npm ci --prefix web
   npm run golden:check --prefix web

3. Open the local web workspace:

   npm run dev --prefix web

4. Inspect the evidence chain in evidence/, parameters/, mission/, and artifacts/.
5. Run the full repository gate with make check.

What this project is

• A reproducible research demo with deterministic simulation in sim/ and browser UI in web/.
• A public evidence chain linking claim -> assumption -> model -> artifact -> source.
• A governance-layered repository with auditable decisions and change rationale in engineering/.
• A small hash-routed workspace with mission, parameter, failure-surface, and optimization views over one tracked browser dataset.
• An artifact-backed Capsule Lab that exposes capsule materials, target/flight-time/velocity choices, and p05/p50/p95 capsule-only survivability rows.
• A Mission Feasibility screen artifact that makes target, velocity, flight years, dust/gas exposure, risk-budget linkage, and energy/cost proxy blockers inspectable.
• A User Mission Run catalog, Runtime Scenario Generation artifact, and strict local pack validator that assign stable run ids, expose run recipes, and execute the Mission DAG into ignored ops/reports/ evidence packs without claiming a server run store.
• A Mission Probability Coupling artifact that factors selected mission probability into open external factors plus a closed capsule/data review proxy backed by compact DAG snapshots.
• An Uncertainty Interactions artifact that exposes pairwise uncertainty residual screens while keeping covariance and correlation evidence external.
• An Evidence Upgrade Campaign artifact that ranks source-review work across all parameter claims while keeping trust promotion and source correctness external.
• An Optimization v2 artifact that adds four-axis Pareto screening over success, risk, qualification gap, and cost proxy without claiming a global optimum, procurement estimate, or hardware qualification.
• A Mission DAG v2 boundary artifact that maps every DAG module to v1 wrapper support, v2 trace/backend requirements, failure taxonomy coverage, and blocked independent-backend claims.
• A full-v2 roadmap closure artifact that tracks all 15 complex next-step items as repo-native contracts while keeping external evidence gaps open and visible.
• An External Proof route and artifact suite that tracks review execution records, repo analytic cross-checks, capsule qualification matrix, exportable reproduction packs, external evidence intake rules, evidence closure, and publication-readiness gates without fabricating third-party validation or lab evidence.
• The public parameter workspace is intentionally limited to mission/design/environment parameters; internal model literals stay in canonical registries but are excluded from the browser-facing index.

What this project is not

• Not a mission-ready flight system.
• Not a full-fidelity physical model of all interstellar hazards.
• Not a substitute for hardware qualification, regulatory certification, or operational safety review.

Scientific Trust Model

• realistic results are policy-facing outputs: no speculative (D-grade) parameter influence is allowed.
• speculative results are exploratory outputs: non-physical controls are allowed, but always explicitly labeled.
• Dual-mode reporting is mandatory for mission runs; realistic and speculative outputs are never merged into a single undisclosed score.

Deterministic Reproducibility Contract

• Simulation core is pure and seeded; no runtime randomness or wall-clock dependence in model math.
• Canonical serialization is stable and checksum-based.
• CI verifies golden output integrity and blocks unauthorized drift.

Reproducibility Contract (short version)

1. Clone repository and install dependencies.
2. Run deterministic checks (make check and npm run golden:check --prefix web).
3. Treat checksum drift as a blocking event unless model evolution protocol is followed.

Reproducibility in 30 seconds

make check
python3 scripts/ci/required_paths_validate.py --strict
npm ci --prefix web
npm run golden:check --prefix web
python3 scripts/mission_baseline_check.py --verify-deterministic

Full contract: REPRODUCIBILITY.md Repository map: docs/README.md

Web Workspace

The web workspace is included in web/ and can be built locally. The hosted demo is published from the same deterministic build gate:

• Live demo: https://ivgolovach.github.io/interstellar-archive/
• Local dev server:

  npm ci --prefix web
  npm run dev --prefix web

Workspace routes:

• #/mission
• #/parameters
• #/failure-surface
• #/optimization
• #/optimization/optv2-pt-000
• #/capsule-lab
• #/mission-feasibility
• #/mission-runs
• #/mission-probability
• #/uncertainty-interactions
• #/evidence-campaign
• #/external-proof

Golden checksum section

• Current golden checksum: bda117f5294cdd3147506d0b3f3f92c99d4113857631230b05bacdd0bd70288c
• Source of truth: sim/golden/golden_checksum.txt
• Determinism status payload: artifacts/determinism_status.json

Architecture link

• ARCHITECTURE.md
• docs/ARTIFACT_POLICY.md

Governance link

• engineering/GOVERNANCE.md
• engineering/DECISIONS.md
• engineering/CHANGELOG.md
• PROJECT_TIMELINE.md
• PUBLICATION_POLICY.md

Public Research Artifacts

• CITATION.cff
• docs/RESEARCH_LICENSE.md
• THREAT_MODEL.md
• MODEL_VERSION.json
• artifacts/research_signals.json

Version + Release section

• Current repository version: see VERSION
• This public repository starts from the 0.3.0 publication-candidate content snapshot.
• Release tags should be created only after local validation, web validation, and remote GitHub checks pass for the exact commit.

License

• Source code is licensed under MIT.
• Research text, schemas, data, and generated artifacts are licensed under CC BY 4.0.