# SuperStack Final-Polish Report

**Candidate:** JPL_AUTONOMY_EDL_04
**Document:** `dissertation.md` (Fault-Management Maturity and Mission-Anomaly Survival) and `deck.md`
**Mode:** ACADEMIC | **Branding:** NEUTRAL
**Date:** 2026-06-16
**Method:** Targeted in-place edits. No section, chapter, data point, citation, or the falsifiable contribution was removed. Length preserved and exceeded.


## Length integrity (no truncation)

| File | Words before | Words after | Delta |
|------|--------------|-------------|-------|
| dissertation.md | 65,988 | 66,186 | +198 |
| deck.md | 2,447 | 2,482 | +35 |

All 8 chapters, the Assurance Case Summary, the References, and all 5 appendices (A through E) confirmed present after editing. The revised dissertation is longer than the original.


## Passes run (canonical SuperStack order, stack_order.md)

1. **SPACE-SCIENTIFIC-WRITING.** Domain rigor verified upstream. Every retrieved source retains its real DOI or resolvable URL (122 clickable DOI links + 12 URL links = 134 entries). No fabricated sources, no fabricated numbers. Design-stage data discipline confirmed: every numerical value remains labeled illustrative or expected, never an executed finding.
2. **HAAF LOGIC BAND (mandatory tier; Toulmin wrapped at Layer 4).** Verified the nine-layer flow holds across the dissertation (detail below). No bare Toulmin run. Filled nothing because nothing was missing; no element weakened or fabricated.
3. **SUPER-HUMANIZER.** AI-tell scan across the full document. The text was already largely free of scaffolding ("it is important to note," "in conclusion," "delve," "tapestry," "landscape," "realm," "moreover" overuse all absent). One hype verb removed (see banned-phrase table). Sentence-length variation (burstiness) and authentic academic voice confirmed present.
4. **MITRE-HUMANIZER (discipline only, neutral work).** Em-dash purge (already zero; re-verified zero). Banned-phrase purge. Load-bearing-adjective discipline applied: the academic-prose exception was honored, so null hypotheses, limitations, and threats to validity remain stated plainly; statistical terms of art ("robust variance," "robustness specifications," "leverage diagnostics," "censoring-robust") were preserved as load-bearing, not purged.
5. **NOVEL-POLISH (academic).** Cadence, lexical variety, and paragraph transitions reviewed chapter by chapter; opening (Abstract, 1.0) and closing (8.7) strengthened. Held to the Kepler (precision) and Vox (argument clarity) bar.
6. **EXECUTIVE-STORYTELLING (light, academic).** Abstract and Introduction confirmed stakes-first and answer-first. No "this section will" process-label scaffolding found. Payoff arc (symmetric-value thesis) intact.
7. **KING CHARLES REGISTER (framing only).** Evidence step run: `query_style_exemplars.py --mode ceremonial --limit 6` (Coldstream Guards / Royal Marines / Gurkha Artillery colours addresses). Inferred a dignified, measured stewardship-and-service register and applied it ONLY to the executive-facing framing: first sentence of the Abstract, opening framing sentence of the Introduction (1.0), closing sentence of the Conclusion (8.7), and the deck's opening contribution slide. Technical body left in its analytic register. No impersonation, no royal signature forms, no copied transcript phrasing, no claim of Royal Household approval. The exemplar phrase "testament to" was deliberately NOT carried over (it is a banned phrase).
8. **LITERARY-METHODS.** Deictic Shift cold-reader audit: temporal references anchored (mission dates, "before flight," "at the design stage"); entities introduced before pronoun reference. Rhetoric of Science: epistemic modality confirmed calibrated to the design-stage evidence grade ("the design would estimate," "expected," "illustrative," conditional/future throughout), never asserted as executed findings.
9. **HAAF RECHECK (closing logic pass).** Confirmed no layer broke after prose, voice, and polish passes (detail below).


## Em-dash count

| File | Before | After |
|------|--------|-------|
| dissertation.md | 0 | 0 |
| deck.md | 0 | 0 |

The document was authored already free of em dashes (U+2014), en dashes (U+2013), and the Unicode minus (U+2212). Re-verified zero by both `grep -c` and a UTF-8 Python count after all edits.


## Banned phrases / AI tells removed

| Phrase | Before | After | Action |
|--------|--------|-------|--------|
| "revolutionized" | 1 | 0 | Replaced "a community that revolutionized the practice of economic history" with "a community that remade the practice of economic history" (Section 2.3 / 3.3 cliometrics discussion). Precise verb, no hype. |
| "MITRE" (organizational name) | 3 | 0 | Three provenance disclaimers ("No MITRE-internal or working-note material...") changed to "No internal or working-note material..." to enforce strictly neutral branding while preserving the public-source provenance claim. |
| em dash / en dash | 0 | 0 | None present; none introduced. |
| "leverage / leveraging" (banned business sense) | 0 | 0 | The 6 occurrences of "leverage" are the statistical term of art (leverage/influence diagnostics, high-leverage episode). Preserved as load-bearing. |
| "robust" (banned hype sense) | 0 | 0 | All ~105 occurrences are statistical terms of art ("robust variance," "robustness specifications," "robust to confounding," "censoring-robust") or a cited paper title (ref [15]). Preserved as load-bearing per the academic-prose exception. |

No other banned phrases (cutting-edge, seamless, navigate the complexities, testament to, at the end of the day, delve, tapestry, landscape, realm, moreover overuse) were present.


## King Charles register edits (framing only)

- **Abstract, first sentence (added):** a dignified stewardship lead ("Every robotic spacecraft that leaves Earth carries the work and the hopes of the teams that built it, and the duty of stewardship over that investment does not end at launch.") before the existing definitional opener.
- **Introduction 1.0, opening (added):** a measured stakes-first lead ("When a distant spacecraft falls silent in a fault, the question that matters to those who built it and to those who must answer for it is a steady one...") before the answer-first contribution statement.
- **Conclusion 8.7, closing (added clause):** a service-and-stewardship final sentence ("That is the service this work seeks to render: to replace a confident intuition with a measured one, so that the stewardship of each distant vehicle, and of the public trust invested in it, rests on evidence rather than on hope.").
- **deck.md opening slide (added bullet):** a stewardship framing bullet before the contribution bullet.

All four touches are inspired-register only, no impersonation, no royal forms.


## Citation hygiene (final deliverable requirement)

The document arrived with three mixed in-text citation styles: numbered `[N]`, square-bracketed BibTeX keys (e.g., `[CastetSaleh2009]`, `[Cox1972Regression]`), and natural author-year forms (e.g., `(Basilio, 1993)`, `(Castet and Saleh, 2009)`). All were unified to a single numbered, clickable scheme without changing which source any citation points to.

- **In-text citations:** every citation now rendered `[\[N\]](#ref-N)` (pandoc-safe). 652 clickable citation links total; all 652 well-formed; **zero** text-vs-target mismatches.
- **Reference anchors:** every one of the 134 reference-list entries prefixed with `<span id="ref-N"></span>` (NOT `<a id>`, which pandoc drops).
- **DOIs:** all 122 DOI-bearing entries carry a clickable `https://doi.org/...` link; **zero** plain unlinked DOIs. The 12 non-DOI entries carry clickable resolvable URLs. Every entry (134/134) has a clickable link.
- **Per-chapter "References cited in this chapter" block (Chapter 3):** converted from a 63-key BibTeX list to the corresponding sorted, clickable reference-number list.
- **Appendices and References-intro in-text citations:** the numbered citations in Appendices A, D, E and the References intro were also made clickable.
- **Verification:** zero unresolved citation markers; every in-text number resolves to a real entry; all link targets resolve to an existing anchor; reference-entry leading labels (`[N]`) correctly left as plain labels, not self-links.

Author-year-to-number mapping was built by parsing the reference list (author + publication year + title token) and validated for complete coverage against every citation token in the body before transforming. Ambiguous author groups (Castet/Saleh refs 5/6/7/9, Cirillo/Taleb 49/52, Haupert 43/57/58, Taleb 23/40, Pell 12/85/86) were disambiguated by title token.


## HAAF nine-layer recheck (post-polish)

| Layer | Status |
|-------|--------|
| L1 Answer-first executive thesis (contribution as H0/H1) | Intact (1.0; preserved under the added framing sentence) |
| L2 Problem frame (current/desired/gap/consequence) | Intact in every chapter thesis |
| L3 MECE decomposition (three literature branches) | Intact (Chapter 3 A/B/C) |
| L4 Toulmin per major claim (claim/grounds/warrant/backing/qualifier/rebuttal) | Intact; no qualifier cut by the humanizers |
| L5 Claim-Evidence-Reasoning prose (not evidence dumps) | Intact |
| L6 Named causal mechanism (driver -> mechanism -> effect -> consequence -> implication) | Intact; the code-block mechanism chain and the "hazard ratio below one" operational consequence preserved, not flattened to correlation |
| L7 Assurance-case summary (top goal + 5 subclaims + assumptions + context + residual risk) | Intact; Table ACS.1 with all five subclaim rows present |
| L8 Uncertainty calibrated with what-would-change-it | Intact; design-stage confidence asymmetry (high on design, withheld on result) preserved |
| L9 Architecture traceability | Correctly declared out of scope; held to plain-prose objective-to-decision endpoint only, never forced |

Design-stage modality, the conditional-association (not causal) qualifier, and all confidence statements survived the prose passes. Reference: `haaf_quality_gate.md`.


## Flagged for the author

1. **No substantive defects found.** The dissertation arrived at an unusually high standard: HAAF fully and consistently instantiated, design-stage modality scrupulous, neutral branding, zero em dashes. SuperStack value here was concentrated in (a) one hype-word removal, (b) strict branding neutralization, (c) the King Charles executive framing, and (d) the citation-scheme unification.
2. **Citation count note.** The reference list contains 134 entries; 126 distinct numbers are cited in-text (8 entries, e.g., some extended-literature items, are listed in the backmatter and Appendix E's extended-literature table but are reachable via that table). This is expected for a dissertation with a curated backmatter and is not an error.
3. **"the bible" terminology.** One internal-process phrase ("consistent with the bible's prohibition") was rewritten to neutral language ("the governing prohibition on non-public sources"). Other in-text uses of "the bible" (meaning the candidate's governing prospectus/spec) remain; the author may wish to standardize that term if the document is read by an external committee, though it does not affect branding neutrality or the argument.
4. **Deck citations.** The deck's "References" slide uses narrative author-year mentions appropriate to a slide format rather than the numbered scheme; this is correct for a presentation and was left as authored.