passon — Mind the Gap proofread integration check
Identity
- Project: Mind the Gap, final proofread / integration check before the final pass.
- Practitioner: Paul Roebuck.
- Instance name: none. This is one-off proofread work, not the Book Man / Jose drafting lineage. Do not adopt either name.
- Session date: 3 June 2026.
Recent state
- Manuscript: V83 per colophon. File
MindTheGap_FinalManuscript_0.docx. Actual word count 57,412.
- What happened in the previous session: full integration check produced against the docx. 22 findings catalogued, grouped by severity (Red / Amber / Bronze) and by category (mechanical / decision / external verification / Word-render check). Workflow agreed: four phases — Decisions → Verifications → Tracked-changes pass → Word visual sweep.
- Status: beginning Phase 1.
- What is locked: nothing. No changes made to the manuscript. Findings catalogued only.
Pending decisions — Phase 1 work
Thirteen substantive decisions plus two external verifications. Grouped by territory.
Body and biography
- Father's age at death. Body says five (four mentions across Chs 3, 14, 14, 15). Foundations table row for Ch 15 says three. Which is canonical?
- Four-rooms list. Chapter 9 (factory / boardroom / therapy / AI). About the Author (factory / boardroom / therapy / cancer ward, with consulting as the fifth). Different lists. Which version stands?
- Tongue cancer vs mouth cancer. Body uses mouth consistently. About the Author uses tongue. Which terminology?
Canonical lines register
- Chapter 4 line wording. Front-matter note quotes "Statistical at mechanism. Commercial at platform. Fluent at every level..." — body line 1115 reads "Statistical at the level of mechanism. Commercial at the level of platform. Fluent at every level..." Which version wins?
- Second canonical line location. Front-matter says appears in Chapter 6; actual is Chapter 7. Update front-matter to Chapter 7? (Confirming.)
- Canonical Lines Register itself. Bibliography refers to "Canonical Line 1" (Bond entry) and "Canonical Line 31" ([name held] entry). No register exists in the book. Include the register, or remove the cross-references?
Acknowledgments and front matter
- NAM identity. Line 99 — broken bold pattern, name appears as placeholder "NAM". Context (CEO at K3 in paragraph that follows) suggests Andy M[…]. Confirm the name.
- Subtitle. Cover (line 3, no full stop, clean markup) vs copyright page (line 19, full stop, broken bold). Which is canonical?
- "see you on the Maasai" — dedication to Kieren. Maasai people or Maasai Mara place? Preserve verbatim or correct?
Author bio and metadata
- Plugs timeline. Chapter 9 line 2001 says "about a year" of AI work. Chapter 13 line 2745 says "two years of serious AI work." Same Plugs episode. Resolve.
- Word count in colophon. "fifty-four-thousand-word manuscript" claimed (line 3469). Actual 57,412. Update to what?
- Author photo. Line 49 placeholder. Supply file or defer to layout phase?
Endorsements
- Reviewer attribution. Three quotes (lines 3485, 3489, 3493) all under Ewan [name held]'s name. Intended as three quotes from one reviewer, or were additional reviewers planned and dropped?
External verifications — Phase 2, parallel
- ISBN registrant codes. Paperback 1-0369, ebook 1-0667143. Different registrants. Verify with Nielsen.
- GPT model name. Cover credits "GPT-5.5" (OpenAI). Verify against current OpenAI public model name.
Deferred to layout phase
- Author photo (decision 12 may push it here)
- Font (line 3471 placeholder)
- Index (line 3459 — compiled post-typeset)
Working method (compressed)
Direct, precise, grounded, unsentimental. No waffle, no platitudes. Frame first; structure follows. Short bold headers where useful. Distinguish known / inferred / speculative — never guess silently. Hold views with conviction; update on new argument, not on pushback alone. Paul decides every meaningful change — propose, do not action.
Flag Protocol: RAG (Red / Amber / Green) × GSB (Platinum / Gold / Silver / Bronze). Default no flag. Used only when adding signal. Red placed by consequence not size.
Voice register for Paul's content: direct, exact, working-class grounded — factory floor → boardroom → therapy room → cancer ward. Preserve coined phrases exactly. No smoothing into corporate / therapeutic / AI register.
First questions for the next session
- Which decision to start with? Recommended: highest-consequence first — father's age at death, since it is the biographical anchor of the framework. Paul's call.
- Scope of Phase 1. All thirteen decisions in one session, or split across two? Estimated 30–45 minutes if decisions are mostly quick; longer if any need thinking time.
File pointers
Upload to the new chat alongside this passon:
MindTheGap_FinalManuscript_0.docx — original manuscript, untouched.
MindTheGap_IntegrationReport.md — full 22-finding catalogue with line references.
- This passon document.
Global memory note: existing Mind the Gap entries reference V42 (~45,500 words). Final manuscript is V83 (~57,400 words). Memory will be updated after Phase 4, not now — work is mid-flight.
Notes
- One-session decision sweep, not long-running drafting. No instance name. No tracker beyond the integration report itself.
- Phase 3 (tracked-changes Word file production) is the heavier phase and warrants its own fresh chat after Phase 1 + 2 complete.
- This work sits outside any Claude Project. No Project Files vector. Manual uploads only.
- Soft footer for the new instance: if anomalies appear on opening the manuscript that the integration report did not flag, raise them silently before Phase 1 begins.