[12-0] Write the final Goals 1-12 findings report

[duckyquang]

Task [12-0] | GEML Goals 6-12 workshop plan | Size: L | Revised: 2026-07-26

Authoritative scope revision: This body supersedes the earlier production-scale wording. Goals 1-5 and their 250k-v1 corpus, EML trees, DAGs, motifs, and reports are immutable inputs: do not regenerate or modify them. The project is targeting a resource-bounded MathNLP workshop submission. Preserve failures, unsupported cases, invalid outputs, and timeouts as explicit rows.

Dependencies

Hard prerequisites: all completed goal summaries/gates, [11-1]/[11-2], and available #82 external-reference results. Missing work must be reported as missing/deferred, never invented.

Objective

Generate one audit-linked final report for Goals 1-12 that separates structural evidence, controlled learned benchmarks, compiler-v2 conformance, external LLM references, limitations, and explicitly deferred scaling.

Exclusive file ownership

Owned paths - edit only these paths:

• docs/goals/FINAL_REPORT.md
• src/geml/analysis/final/report.py
• tests/analysis/test_final_report.py

Read-only dependencies:

• All final goal summaries, gates, run manifests, and artifact indexes under outputs/final/**

Tasks

• Validate source manifests/checksums and generate traceable tables with source paths/config hashes.
• Summarize Goals 1-5 without changing or rerunning them; distinguish structure/compression from downstream predictive utility.
• Report Goals 6-9 controlled results with denominators, uncertainty, compute, failures, and OOD limits.
• Report Goal 10 only as opt-in inverse-trig/constant compiler conformance; state that no corpus/learning rerun was done.
• Report Goal 11 fixed-scale efficiency and LLM external context; explicitly exclude 10-100x scaling conclusions.
• Include threats to validity, null/negative results, resource budget, reproducibility status, and a claim-to-artifact index.

Acceptance criteria

• Every numerical claim links to a frozen artifact row/checksum.
• No missing, failed, or deferred experiment is represented as complete.
• External LLM results are labeled non-controlled and kept outside controlled gates.
• The report clearly states the fixed 250k-v1 evidence boundary and no scaling-law claim.

Validation

• python -m pytest tests/analysis/test_final_report.py
• Regenerate the report from frozen manifests and verify links/checksums.
• Run the repository-wide standard validation commands.

No individual test, lint, or smoke command may exceed approximately 30 minutes. Production experiment cells may run longer only when sharded, resumable, checkpointed, and independently auditable.

Out of scope

• New experiments, result repair, corpus regeneration, speculative conclusions, or paper-specific marketing language.

PR checklist

• I edited only the owned paths above and did not change another issue's contracts.
• I used only current repository specifications, the complete issue, authoritative public mathematical sources, and official library documentation.
• Tests use tiny hand-written or temporary fixtures and do not require outputs/ or production artifacts.
• I retained and reported every failure, timeout, unsupported input, and validation error.
• I recorded configuration hashes, exact seeds, commit/package/hardware metadata, and any scientific assumption or metric change.
• I ran python -m pytest, python -m ruff check ., and python -m ruff format . --check before handoff.

Clean-room rule: Do not inspect or reuse the prototype repository, its code, tests, schemas, helpers, architecture, or commit history. Stop and ask when a shared or scientific interface is ambiguous.