TS — TFW-55 / Phase A: Project North Star Essay¶

Date: 2026-08-26 Author: Coordinator Status: 🟡 TS_DRAFT — approved by the owner 2026-08-26; ready for /tfw-handoff Parent HL: HL-TFW-55 Phase HL: Phase A derivation Master freeze: a60bc6d Research: Iteration 1 · Iteration 2

1. Objective¶

Rewrite .tfw/README.md as a coherent English Project North Star essay of at most 2,000 words. It must lead a reader from the problem of invisible, discontinuous AI-delegated work to the Philosophy of Trace, define TFW as the methodology built on that philosophy, designate citable purpose/principles/non-goals, and route readers unambiguously to current mechanics, history/evidence, and the future BoK without creating that BoK.

2. Scope¶

In Scope¶

Replace the current implementation-led philosophy paper with the approved problem-led narrative derived from mini-essay 0.8.
Define Philosophy of Trace → TFW methodology → realizations/workflows → this self-applying reference implementation.
Explain selected Trace, continuability, Questions, the human/agent authority boundary, operational self-awareness, and risk-proportional assurance in ordinary language.
Add visible NS1 — Purpose, NS2 — Principles, and NS3 — Non-goals anchors inside .tfw/README.md.
State the North Star / future approved BoK / living specification / corpus authority model.
Remove or qualify the overclaims listed in master HL §2 and Phase HL Deliverable 6.
Preserve source, claim, word-count, link, and subtraction checks in the phase evidence file and RF.

Out of Scope¶

Root README.md, README.ru.md, README.kk.md, and Task Board content beyond normal planning status updates.
Creation, publication, structure design, or versioning of the BoK.
Framework mechanics: conventions, glossary, workflows, templates, Editions, adapters, configuration, and runtime.
Methodical guide, book, course, university, legal/IP, market, certification, product, and brand work.
Claims of proven human comprehension, learning, transfer, adoption, category novelty, or component novelty.

3. Principles Check¶

#	Principle from master HL §7	Enforced by	Gate
P1	One North Star, several depths	AC-4, AC-7	One purpose/principle authority; mechanics and history remain elsewhere; future BoK is subordinate
P2	Philosophy before machinery	AC-1, AC-8	Problem-led opening and heading order; no lifecycle-first exposition
P3	Human purpose remains human	AC-3	Every authority/accountability sentence inspected
P4	Trace, not transcript	AC-2	Selected material continuity plus explicit exclusions
P5	Self-awareness must be operational	AC-6	Six capabilities plus non-consciousness/non-authority limit
P6	Subtract before adding	AC-8, AC-9	Unchanged word ceiling plus removed/moved/added ledger
P7	Provenance before polish	AC-5, AC-9	Source and claim-state classification in evidence
P8	Experience reveals the system	AC-6	Editions answer visible coordination/risk needs without universal pedagogy claims
P9	Refutation before canonicalization	AC-5, AC-9	Strong framework-only explanation and negative findings constrain wording

4. Affected Files¶

File	Action	Description
`.tfw/README.md`	MODIFY	English Project North Star and canonical foundation essay
`phase-a/evidence/EV__phase-a__canonical_foundation_essay.md`	CREATE	Per-AC evidence, source/claim/subtraction ledger, word/link/anchor checks

Budget: 2 affected files; 1 production modification; 1 new evidence file; estimated ≤600 changed lines. This is within project limits (30 files, 15 new files, 3,000 LOC, 30 modified files) and does not increase the 2,000-word essay ceiling.

5. Required Reading and Source Handling¶

Read in this order:

Master HL frozen at a60bc6d, especially §§1, 3–7 and Amendments A2–A5.
Phase A HL.
All nine files in C:\Users\c0rpa\obsidian\c0rp\c0rp\Книга TFW\мини-эссе 0.8\, in numeric order.
C:\Users\c0rpa\obsidian\c0rp\c0rp\Книга TFW\body of knowledge\00. Архитектура TFW.md.
Iteration 2 RES, then Iteration 1 RES; use F1–F4/R1 as private completeness checks only.
Current .tfw/README.md as rewrite/subtraction material.
body of knowledge/version 0.1.md only as the final completeness, contradiction, hypothesis, and risk check.

The executor must not treat source recency, founder authorship, or fluent phrasing as evidence of universality. If a source claim conflicts with master boundaries, narrow it in prose and record the disposition in EV.

6. Acceptance Criteria¶

AC-1: Problem-led definition and architecture¶

The essay earns one functional definition through the problem before it explains machinery.

[ ] The opening makes invisible/discontinuous AI-delegated intellectual work concrete before naming framework artifacts.
[ ] One concise definition identifies Philosophy of Trace as foundation, TFW as methodology, and workflows/implementations as realizations.
[ ] The repository is a self-applying reference implementation, not the only realization or proof that every historical file is normative.
[ ] Methodology is presented as the owner-selected position supported by function, composition, and exclusions—not as a taxonomy or novelty proven by research.
[ ] The essay does not begin with lifecycle stages, artifact names, installation, or code regeneration.

Gate: EV maps the opening and definition to these five units and records the exact section/anchor for each.

Evidence: N/A — this is a document-structure and content criterion; reproducible inspection belongs to the Gate and required EV file, not a live environment.

AC-2: Trace and continuability¶

The essay distinguishes selected material continuity from chat history and from automatic truth.

[ ] Trace is a selected, durable, inspectable material record rather than raw transcript, hidden chain-of-thought, complete memory, or guaranteed truth.
[ ] Ordinary language covers the internal F1–F4 floor: human purpose/authority; bounded delegation; selected material trace plus result/current state; continuation or explicit close.
[ ] Goal, Value, Question, Task, Context, and Boundaries are introduced only to the extent needed to show orientation, uncertainty, bounded work, and drift control.
[ ] A “good Trace” enables inspection, reconstruction, comparison, and continuation; it does not promise identical deterministic reproduction.
[ ] Evidence, review, and verified knowledge are risk-proportional assurance layers, not universal minimum artifacts.

Gate: EV maps each F1–F4/R1 semantic unit to public prose without exposing those labels in the essay.

Evidence: N/A — this is a semantic document criterion; the required EV mapping provides the reproducible inspection record.

AC-3: Human/agent authority boundary¶

Delegation remains useful without transferring legitimate human responsibility to an agent.

[ ] Human purpose, legitimate authority, judgment, acceptance, accountability, and stop responsibility are explicit.
[ ] Agents receive bounded work with purpose/context, scope/boundaries, and acceptance authority.
[ ] The essay does not require complete domain mastery before delegation; it requires enough understanding to define purpose, bounds, acceptance, authority, and escalation.
[ ] Observability is necessary for inspection and correction but is not presented as a guarantee of correctness.
[ ] Ownership of organizational memory is described contextually; employee, employer, client, privacy, security, and IP interests are not erased by an absolute slogan.

Gate: EV records the exact authority-bearing sentences and a disposition for the three narrowed source claims: delegation/understanding, observability/correctness, and memory ownership.

Evidence: N/A — this phase makes no live authority or human-behavior claim; sentence inspection and claim dispositions are recorded in EV.

AC-4: Citable Project North Star¶

The essay contains a usable North Star rather than a decorative heading.

[ ] Visible Markdown anchors/headings designate NS1 — Purpose, NS2 — Principles, and NS3 — Non-goals.
[ ] NS1 states why TFW exists and what value it protects when intellectual work is delegated.
[ ] NS2 contains a small stable principle set; every principle changes an observable work choice.
[ ] NS3 explicitly rejects becoming: a prompt/chat archive; deterministic generator; replacement for human judgment; maximum-documentation bureaucracy; vendor-bound tool; claim of untested capability.
[ ] A coordinator can use NS1–NS3 to reject an excessive but goal-compatible proposal.
[ ] North Star additions fit the existing 2,000-word essay ceiling; no limit increase is requested.

Gate: EV contains an anchor/link check and two short purpose tests—one aligned proposal and one excessive proposal—with a deterministic accept/reject explanation from NS1–NS3.

Evidence: N/A — anchors, links, and purpose tests are reproducible local document checks recorded in EV.

AC-5: Claims remain evidence-bounded¶

The rewrite does not replace technical exaggeration with philosophical exaggeration.

[ ] No category/component novelty claim is made; adjacent provenance, ADR, docs-as-code, oversight, and knowledge practices are not claimed as TFW inventions.
[ ] Traces Over Code does not imply disposable code, deterministic regeneration, or a software-only method.
[ ] “same lifecycle / same artifacts,” “documentation maintains itself,” deterministic output, provenance-only authority, and unbounded “AI team member” claims are removed or qualified.
[ ] Mini-essay claims about reproducibility, delegation expertise, observability, and memory ownership use the bounds in AC-2/AC-3.
[ ] The Taylorism comparison is omitted; it is unnecessary to the short North Star argument.
[ ] Owner philosophy, teaching observation, project evidence, inference, and open hypothesis are not collapsed into one factual voice.

Gate: EV has one disposition row for every listed old/new overclaim: removed, qualified, or retained with source and reason.

Evidence: N/A — evidence-bounded wording is verified through source accounting and document inspection in EV, not through a live environment.

AC-6: Self-awareness and Editions remain bounded¶

The essay explains capability and proportional implementation without anthropomorphism or pedagogical doctrine.

[ ] A self-aware project can answer the six questions in master HL §3 from durable artifacts.
[ ] The term implies neither sentience, self-generated authority, nor independent accountability.
[ ] Light contains the semantic continuity floor; Assisted and Full add coordination/assurance as risk requires.
[ ] Light → Assisted → Full is one observed founder-led, problem-led route, not a universal maturity law or proven superior pedagogy.
[ ] No retention, transfer, instructor-independence, adoption, or field-effect claim is made.

Gate: EV maps the six questions and records the Editions paragraph against Iteration 2 evidence limits.

Evidence: N/A — this criterion explicitly avoids field-effect or comprehension claims; the bounded terminology check is recorded in EV.

AC-7: Authority and update paths are unambiguous¶

The essay tells a reader where each kind of truth belongs.

[ ] Project North Star owns stable purpose, principles, boundaries, and non-goals.
[ ] Conventions, glossary, workflows, templates, and Editions own current implementation mechanics.
[ ] Task traces, verified knowledge, and Git history form the primary history/evidence corpus.
[ ] A future approved, versioned BoK may expand concepts/hypotheses/evidence limits but remains subordinate to the North Star and is explicitly a separate task.
[ ] The working Obsidian BoK is never linked or described as current repository authority.
[ ] No new canon manifest, claims register, or governance surface is created or proposed.

Gate: EV resolves four scenarios to one destination each: purpose conflict, mechanics update, historical “why,” and future concept/hypothesis expansion.

Evidence: N/A — authority routing is a document-contract criterion verified by scenario inspection in EV.

AC-8: The essay is concise, coherent, and navigable¶

The result reads in one sitting and sends detail to its proper owner.

[ ] English prose is ≤2,000 whitespace-delimited words.
[ ] The narrative follows the Phase HL problem-led map or RF documents a clearer equivalent ordering and reader-value reason.
[ ] No task status table, exhaustive artifact catalog, exact tool walkthrough, installation guide, changelog, or BoK outline appears.
[ ] Every abstract term connects to observable behavior, authority, or project capability.
[ ] A reader can reach current mechanics, a practical Editions/start route, and the project corpus/history.
[ ] Every local Markdown link resolves; NS1, NS2, and NS3 are directly linkable.
[ ] Phase B can derive English, Russian, and Kazakh doorways without inventing a second philosophy.

Gate: Automated word/link/anchor checks plus manual heading-order and content-role review; before/after words and lines recorded in EV.

Evidence: N/A — word, link, anchor, and navigation checks are reproducible local inspections; no deployed or live environment is claimed.

AC-9: Provenance and subtraction are reviewable¶

Normal TFW traces show what entered, left, narrowed, and remained uncertain without creating a public claims registry.

[ ] EV maps every AC to exact essay sections and verification results.
[ ] EV includes a compact source/claim ledger: source class, source location, old or candidate claim, disposition, new bounded claim, and evidence limit.
[ ] EV records current/final word and line counts, net change, and removed/moved/added content classes.
[ ] EV records the BoK v0.1 completeness check while keeping all unresolved hypotheses/open questions non-canonical.
[ ] No separate public provenance or authority artifact is created.
[ ] EV contains no placeholders; every status agrees with the TS gate result before RF claims completion.

Gate: EV has per-AC verdicts, all required ledgers/checks, zero placeholders, zero unresolved local links, and internally consistent statuses.

Evidence: N/A — this criterion verifies the phase trace itself through local artifact inspection; EV carries the complete record.

Evidence Artifact¶

File	Description
`phase-a/evidence/EV__phase-a__canonical_foundation_essay.md`	Required structured evidence with per-AC verdicts, semantic/source maps, claim dispositions, North Star purpose tests, word/link/anchor checks, authority scenarios, and subtraction ledger

Phase A makes no live-environment or human-comprehension claim. Its evidence is document inspection, source accounting, completed research bounds, and reproducible local checks.

7. Technical Guidance¶

Treat the Phase HL narrative map as a content skeleton, not mandatory final headings or sentence wording.
Use the mini-essays for the human explanatory path; use F1–F4/R1 privately to test completeness, never as reader-facing jargon.
Use methodology consistently with master A3. Let the function and observable composition do the explanatory work.
Use current .tfw/README.md as rewrite material, not a structure to preserve. Link detailed mechanics instead of abbreviating them.
Keep NS1, NS2, and NS3 stable and directly linkable; headings may add ordinary-language subtitles but not change the IDs.
Do not edit a specification file to make an essay sentence true. Qualify the prose or record a follow-up finding.
Keep examples domain-neutral. A software example may illustrate but may not define TFW.
Do not use long quotations from the owner corpus. Synthesize, preserve provenance in EV, and retain the owner's conceptual voice without copying a draft book into the README.

8. Definition of Failure¶

❌ The essay starts with Full mechanics, artifact names, or installation rather than the reader's discontinuity problem.
❌ TFW is called novel or categorically superior to a framework as if research established it.
❌ Philosophy of Trace, TFW methodology, and a concrete realization collapse into synonyms.
❌ Trace means transcript, hidden reasoning, provenance alone, complete memory, or deterministic reproduction.
❌ Human authority/accountability becomes implicit while agents are described as independent responsible actors.
❌ NS1–NS3 are absent, decorative, uncitable, or unable to reject an excessive proposal.
❌ The working BoK becomes authority, a deliverable, or a public link in this phase.
❌ self-aware is anthropomorphic or Editions become a universal maturity/learning ladder.
❌ The essay exceeds 2,000 words, breaks local links/anchors, or requires another authority surface.
❌ Any framework mechanics, root README prose/translations, course/book material, or brand design is changed in this phase.
❌ EV cannot account for a source claim or overclaim disposition, an AC gate fails, or EV/RF statuses disagree.

9. Phase Risks¶

Risk	Mitigation
The definition becomes dense and unreadable	Reach it through the problem; allow a short paragraph while keeping one primary formulation
Nine essays become nine compressed sections	Consolidate by reader question and enforce one job per section plus 2,000-word ceiling
North Star crowds out the narrative	Use three compact citable anchors and fund them by removing old mechanics/duplication
Evidence ledger becomes a second canon	Keep it phase-local in EV; public essay contains selected meaning only
BoK completeness checking expands scope	Check for omissions/contradictions only; defer structure, versioning, and publication to the next task
Phase B translates draft wording	Phase B remains blocked until Phase A RF and REVIEW are complete

10. Cross-Phase and Follow-up Boundaries¶

Phase B reads the reviewed .tfw/README.md and modifies README.md, README.ru.md, and README.kk.md; it derives all three doorways from the same NS1–NS3 semantics. It does not reopen the philosophy.

The BoK is a separate next task after Phase A REVIEW. It may use the working v0.1 corpus, but must reconcile it with the reviewed North Star and repository evidence under its own HL/TS. Phase A neither pre-approves the draft BoK nor defines that task's detailed structure.

TS — TFW-55 / Phase A: Project North Star Essay | 2026-08-26