Produced 3 June 2026. Read against MindTheGap_FinalManuscript_0.docx (V83 per colophon, 57,412 words). Line numbers refer to the pandoc text extraction; they map cleanly to the docx if you scan section by section.
Findings grouped by consequence. Flag Protocol applied — RAG × GSB. Default no flag. Used only when adding signal.
Line 43 (note on authorship) quotes the canonical co-authored sentence as:
Statistical at mechanism. Commercial at platform. Fluent at every level the user encounters.
The actual line in Chapter 4, line 1115, reads:
Statistical at the level of mechanism. Commercial at the level of platform. Fluent at every level the user encounters.
Two of three clauses differ. The version on line 43 is tighter; the body version is longer. For a book that argues the discipline of precise attribution, the front-matter quotation of its own canonical line must match the line. Either rewrite the body to the front-matter version, or rewrite the front-matter quote to the body version.
Line 43 says the second canonical line "appears in Chapter 6 and again at the close." Actual locations are Chapter 7 (line 1517) and the close (line 2949 / Chapter 15). The first appearance is Chapter 7, not Chapter 6.
Body, four mentions, all consistent: father died when Paul was five.
Back-matter Foundations table, row for Chapter 15, "First wound":
The originating wound of the framework is the most consequential biographical fact in the book. Two different ages are stated. One needs to win.
Every other chapter heading is Heading 1 (renders as # CHAPTER N in extraction). Chapter 8 alone is bold text only (CHAPTER EIGHT, line 1627). Consequences: will not appear in an auto-generated table of contents, will be visually inconsistent with the other fourteen chapter openings in print, any post-typeset Word workflows (running heads, indexer pull) will skip it. Apply Heading 1 style to bring it into the structure.
Line 99: Then *here is *NAM, — bold markup is broken and "NAM" appears in caps. Context: preceding people are introduced as "Bill O.", "Jean B.", "Bob M." and the paragraph that follows is about Andy, a CEO at K3. "NAM" reads as either a placeholder ("name") never replaced, or "Andy M." with the initials run together and the markup shattered. Sitting on the acknowledgments page, the kind of thing the dedicatee will notice first.
Line 3469 (colophon): "this fifty-four-thousand-word manuscript."
Actual: 57,412 words by pandoc extraction. Closer to "fifty-seven-thousand" or "around fifty-five thousand."
Front-matter dedication to Gem (line 215): "over fifty-thousand words" — still true, no change needed there.
Line 3555 (the literal last line of the manuscript): Mind the Gap · CHAPTER FIFTEEN · 1 Back to Contents — Word page footer or TOC navigation element that has spilled into the document body and was not stripped. Should not appear in final manuscript. Check the docx in Print Layout.
Author photograph: [PLACEHOLDER — to be confirmed at layout][PLACEHOLDER] (Index — acceptable, produced from typeset pages)Typeset in [PLACEHOLDER FONT, to be confirmed at layout].All visible in the manuscript as authored text. Final values needed before KDP, or removed if front-matter design doesn't include them.
About the Author replaces AI room with cancer ward and adds the consulting room as a fifth. Both versions describe Paul's life accurately. They cannot both describe the structure of the book at the same time.
(Separately, Chapter 3 line 769 says "Five rooms. One instrument." where rooms refers to domains: Therapy / Commerce / AI / Photography / Grief. Different framing, defensible inside its own chapter, but adds noise if a reader audits the room count across the book.)
Both technically accurate (the tongue is in the mouth) but body uses one term and bio uses another. Mouth Cancer Foundation Ambassador framing sits with mouth.
Defensible if "by then" in Ch 13 means a later point (the Facebook post may post-date the first Plugs use). But to a careful reader, same episode looks dated twice.
Worth deciding canonical subtitle and applying consistently.
Lines 3483–3493: Ewan [name held] credited at top, then three blockquote paragraphs follow. Reads as three quotes from one reviewer — unusual for an endorsement page. Confirm intent.
Cover and copyright pages cite Claude (Anthropic, Opus 4.7) and ChatGPT (OpenAI, GPT-5.5). Opus 4.7 is consistent with current Claude. GPT-5.5 — needs verification against current OpenAI public model name. Getting the collaborator's product name wrong on the cover is the kind of error external reviewers pick up.
The book does not contain a Canonical Lines Register. The Study holds canonical lines as documents. Either include a brief Register in back matter, or remove the "(Canonical Line N)" parentheticals from the bibliography.
Both valid ISBN-13 formats but registrant block lengths differ. Two different Nielsen blocks. Probably intentional if bought separately. Verify with Nielsen before file goes to print.
A consistent pattern shows up where Word bold/italic runs got split across spaces. These may render correctly in Word itself, but each one is a place where formatting is at risk. Cluster in:
* That Gap*. — leading space inside italic.Paul Roebuck, Author, **September 2017. — broken bold inside italic.REVIEWS — should be REVIEWS.BACKCOVER — should be BACK COVER (also missing space).Recommend opening Word doc to Print Layout and visually scanning each. If they render correctly, leave them. If any show stray bold or unwanted formatting, fix.
Already noted at §13. Deliberate single-endorser frontispiece, or additional endorsers planned and not yet added?
Different uses of room. Both work in their own contexts. Worth deciding whether a single sentence in Chapter 3 should distinguish rooms (domains) from rooms (workplaces).
Chapter 8 opens with "Chapter Seven mapped fluency as the surface that conceals what is underneath" (line 1633). Cross-reference correct. Chapter 7 close (line 1625) hands off cleanly. Mentioning only because the heading-level break (Red §4) means a reader of the printed book won't see a chapter division at this transition without the formatting fix.
The book holds together at the level of timeline and biographical spine. Where it breaks is in the metadata around the body — canonical lines register, rooms list in the bio, structural style of one chapter heading, and the placeholders. Body is in better shape than back matter and cover apparatus.