ed/manual_slop
Private
Public Access
0
0
Fork 0
Code Activity

conductor(deob_phases): Initialize 3 phase child spec scaffolds

Browse Source 
Each child spec is lightweight (~120 lines): references the umbrella, gives the deliverable structure, specifies the inputs/outputs, and the 5-phase pipeline.

Phase 1 (lexicon): refines the warmup's draft into a codified operational spec (lexicon.md + terms_catalog.md + dedup_map.md)
Phase 2 (pilot): applies the lexicon to 2 Pass 1 reports (cs229_building_llms + entropy_epiplexity), captures refinements in pilot_report.md
Phase 3 (apply): applies the refined lexicon to 10 remaining Pass 1 reports + 1 cross-cutting synthesis, final apply_report.md

3-layer deliverable per video: translation (side-by-side) + replacement (re-encoded) + decoder (per-term etymology + form anchor + definition history)
4 verification criteria: lossless, bounded, constructively typed, etymology-cited
This commit is contained in:
ed
2026-06-23 00:08:23 -04:00
parent f830798822
commit 256af96bf3
3 changed files with 308 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
conductor/tracks/video_analysis_deob_lexicon_20260621/spec.md
+104
View File
@@ -0,0 +1,104 @@
# Track: Video Analysis De-obfuscation — Lexicon Refinement (2026-06-21)
**Status:** Not started (umbrella + warmup published 2026-06-21)
**Type:** Research-only child track (Pass 2 of 3, Phase 1 of 3 within Pass 2)
**Owner:** Tier 2 Tech Lead (execution)
**Priority:** A (user-blocking)
> **Parent:** This is child #1 of the [video_analysis_deob_20260621](../../video_analysis_deob_20260621/) umbrella (Pass 2 of the 3-pass research campaign). See [umbrella spec.md](../../video_analysis_deob_20260621/spec.md) for the full design.
> **Upstream consumer of:** [video_analysis_deob_warmup_20260621](../../video_analysis_deob_warmup_20260621/) (the warmup produces the initial `report.md` + `prompt_template.md`).
> **Downstream consumer:** [video_analysis_deob_pilot_20260621](../../video_analysis_deob_pilot_20260621/) (consumes the refined lexicon).
---
## 1. Inputs
| Input | Source |
|---|---|
| Warmup `report.md` | [video_analysis_deob_warmup_20260621/report.md](../../video_analysis_deob_warmup_20260621/report.md) |
| Warmup `prompt_template.md` | [video_analysis_deob_warmup_20260621/prompt_template.md](../../video_analysis_deob_warmup_20260621/prompt_template.md) |
## 2. Deliverables
| Artifact | Path | Target |
|---|---|---|
| Refined lexicon | `lexicon.md` | The codified operational spec; the "contract" between the warmup and the apply phases |
| Machine-readable terms | `terms_catalog.md` | The lexicon in a structured form (per-term table with: tier, original notation, re-encoded form, form anchor, etymology, definition history) |
| Dedup maps | `dedup_map.md` | The 3+ noise-dedup maps from the warmup, refined with examples |
| Test cases | (5-10 example transformations, embedded in `lexicon.md`) | Each test case: original notation → re-encoded form, with the lexicon terms used |
**Optional (added per child track execution convention):** `plan.md`, `metadata.json`, `state.toml`.
## 3. Pipeline (5 phases)
Per the umbrella spec §5 (Phase 2 of the umbrella). Each phase commits atomically.
- [ ] **Phase 1: Init.** Initialize the child track (`plan.md` + `metadata.json` + `state.toml` per Tier 2 conventions).
- [ ] **Phase 2: Refine.** Tier 3 worker reads the warmup's `report.md` + `prompt_template.md`. Refines the lexicon by:
- Adding 5-10 test cases drawn from the user's samples
- Adding the "form anchor" requirement to each term
- Adding cross-references to the warmup's report sections
- Formalizing the 4 verification criteria as part of the lexicon spec
- [ ] **Phase 3: Codify.** Produce `lexicon.md` (the operational spec) + `terms_catalog.md` (machine-readable) + `dedup_map.md` (refined maps).
- [ ] **Phase 4: User review.** User reviews the refined lexicon. Approves before the pilot child can start.
- [ ] **Phase 5: Verification.** Idempotency check (the lexicon is consistent across multiple reads), audit, end-of-track report.
## 4. Lexicon structure (what `lexicon.md` contains)
```markdown
# De-obfuscation Lexicon (v1, refined from warmup)
## 1. The 4 Rules (carried from prompt template)
- 1.1 Boundedness
- 1.2 Form anchor
- 1.3 Etymology
- 1.4 Lossless
## 2. The Tiered Terms
- 2.1 Tier 1: Core concepts (set→kind, ∀→forall, ∃→exists, etc.)
- 2.2 Tier 2: Data-oriented pipeline (filter, map, fold, etc.)
- 2.3 Tier 3: Type-theoretic primitives (proof, program, type, etc.)
- 2.4 Tier 4: AI-fuzzing tolerance (try, recover, sandbox, didyoumean, etc.)
## 3. The 3+ Noise-Dedup Maps (refined)
- 3.1 Proofs = Programs = Computations
- 3.2 Sets = Kinds = Types
- 3.3 Functions = Procedures = Words
- 3.4 (additional maps discovered)
## 4. Test Cases (5-10)
- Each: original → re-encoded, with the lexicon terms used
- Drawn from the user's samples
## 5. Form-Anchor Requirement (formal)
- Every re-encoding must have a form anchor
- Definition: a bounded form that the re-encoding projects from the indefinite
## 6. Etymology Requirement (formal)
- Every new term: 1-line origin + 1-line definition history
- Examples from the user's samples
## 7. Cross-References to Warmup Report
- Section mappings: warmup §3 → lexicon §2, etc.
```
## 5. Connections
- **Upstream:** consumes the warmup's `report.md` + `prompt_template.md`
- **Downstream:** the pilot child consumes `lexicon.md` + `terms_catalog.md` + `dedup_map.md`
## 6. Verification
- [ ] All 3 deliverables present
- [ ] `lexicon.md` has the 4 rules + tiered terms + dedup maps + test cases
- [ ] At least 5 test cases included (drawn from the user's samples)
- [ ] User has reviewed and approved
- [ ] `state.toml` updated to `status = "completed"`
## 7. See also
- [Umbrella spec.md](../../video_analysis_deob_20260621/spec.md)
- [Umbrella README.md](../../video_analysis_deob_20260621/README.md)
- [Warmup spec.md](../../video_analysis_deob_warmup_20260621/spec.md)