# Four-role author desk critique · 30 September 2026

**Status:** author simulations of teacher, student, parent/carer and potential funder only. No independent classroom/participant review, actual event/family/account/permission/source record, marker-use/viewer outcome or physical tactile construction. Roles trace source relationships; they aren't validation of interest, consent, learning, accessibility or demand.

**Teacher role:** I followed ten scripts on screen: child chooses source pocket, unfolds event/marking/plan trace, attaches own significance and sharing sources, fresh check, chooses plain/no-marker guide, preserves/repairs one phrase, gives authorised source kit a useful private job, returns/hands over copies with bounded narrative and fresh check. Full actual preflight verifies an actually marked significant event, source/date precision/link, willing account, source-specific sharing and private family supported/open breadth, culturally appropriate scope, marker/audience/return permission. Eight ordinary branches are runnable; complete paper route needs no family contribution/recipient/real event. Seven-minute making requires pre-made folds/flat adult-directed routes. Local authenticity/significance/curriculum/pacing still requires educator judgement.

**Student role:** In author desk simulation I could choose broad tab, source bookmark, flat pocket, oral guide or direct adult/no-touch equivalent. I could ask source questions without telling my family story or explaining a pass. Reported book completion/marking differed from future label plan and today's paper making. Kemi's own reason and Arlo's cousin-sharing were separate source accounts, not everyone. I could preserve first words, repair a sourced phrase and leave absent permission/viewer fields open while keeping useful kit. Fresh B's dated plan/undated drawing didn't establish completed commemoration. These are author interpretations, not learner comprehension/interest/independence. No physical learner/tactile test.

**Parent/carer role:** I checked ten free optional paper/spoken bridges with no purchase/device/printer/outing, returned score/evidence, family interview/private date/photo/belief/relationship/history or public celebration act. Skipping has no assessment cost. Genuine family material only through voluntary child/family-approved private school route; school/others evidence can't silently become family breadth. Silence/withdrawal accepted. No parent/carer reviewed pack. Any translation requires fluent local review.

**Potential funder role:** I traced ten timed scripts/30 routes/20 worked domains/10 bridges/two held fresh checks/key/nine original pairs/exact alternatives, runnable conditional K02 route, exact primary rows/rights and hashes. Stronger actual-event opportunity extends earlier fictional archive but is no actual K02 outcome or full annual coverage. Dated routine, invitation, kit-making and simulated viewer aren't authentic evidence. No time saving, learning gain, demand, universal access, cultural certification or commercial suitability shown. K01/K03/K04 distinct actual-source holds.

**Residual critique:** event versus actual marking/link needs explicit source language, not object interpretation. School/community evidence may show others but leave family breadth open. Particular-person significance isn't a universal celebration meaning. Child choice/no-marker/pass must stay useful. Sensitive or culturally held event material requires appropriate recognised authority; paper route remains full alternative. Don't stage a new celebration to close gaps. PDFs untagged; individual print/screen-reader/tactile/AAC/language/pacing and independent classroom review pending.

## Portable reproduction and checks

From this pack directory: `python3 verify_pack.py` is offline/read-only; `python3 verify_pack.py --negative-checks` rejects in-memory drifts and restores normal inputs. `python3 print/generate_print.py --check-reproduction` writes only temporary storage and compares all 19 print artifacts while preserving pack bytes/size/mtime. `python3 print/generate_print.py --regenerate` deliberately rewrites nine named SVG/PDF pairs/exact alternatives. `python3 verify_pack.py --manifest-json` prints a proposed complete hash receipt for intentional author saving. Python 3/Pillow/DejaVu, librsvg, pdfinfo, pdftotext and pdffonts required. Fixed PDF build clock is metadata, not event time. Published instructions use relative commands; no private workstation path.

## Technical receipt · 30 September 2026

- **Default offline audit:** PASS — nine exact direct workbook XML rows/imported framework/crosswalk; current QCAA alignment/sequence pins; ten 25-minute scripts/eight ordinary actual-event branches; conditional K02/no actual outcomes; 30 routes/20 worked domains/10 bridges; two held fresh checks/key; nine original A4 pairs/full exact alternatives/searchable embedded fonts/measured page and timeline-panel text bounds/keyed event/marking/plan/undated geometry; 55 local links and 37 SHA-256 files (38 total with manifest).
- **Reproduction:** supplied temporary-storage mode reproduced nine SVGs/nine PDFs/exact alternatives, 19/19 byte-identical. All pack bytes/sizes/modification times preserved. No build clock interpreted as event date.
- **Read-only negative audit:** supplied mode rejected 29/29 in-memory drifts: workbook digest, official description, K01 hold, K03 hold, blanket K02 hold, invented actual outcome, timing, ordinary actual-branch removal, actual completed marking/trace removal, particular-person willingness removal, source-specific sharing removal, unsupported family breadth closure, missing no-actual-evidence limit, silence as consent, removed no-marker choice/copied-page return, fresh speaker/key leakage, pace treated as attainment, peer answer as individual evidence, exact alternative/source quotation drift, QCAA sequence pin, planned commemoration changed to reported, square page outside frame, oversized timeline-panel text, private workstation path and stale receipt. Normal audit restored; no bad fixture written; all bytes/sizes/modification times preserved.
- **PDF review:** all nine PDFs rendered/visually inspected together. Event/book/banner, own account/sharing, plan/request, fold/pocket and blank actual fields visible. Initial fresh timeline dates crowded their panels; labels reduced, panel-specific measured bounds added, both fresh pages regenerated and individually re-rendered/inspected. Final source diagrams/text fit without clipping/overlap. Untagged PDFs/full exact no-touch/tactile guidance are not PDF/UA or learner access certification; no physical tactile build tested.
- **Primary evidence:** fresh live ACARA workbook HTTP 200/2,049,917 bytes/pinned SHA matched for this block. Current QCAA Prep alignment direct browser open and P–6 successful official result reopen compared all nine descriptions; sequence direct URL restriction recorded. No direct PDF download/new full HASS entry fetch claimed. ACARA terms rely on earlier successful same-session review, not fresh W37–38 fetch.
- **Code/runtime:** both Python tools pass formatter/lint checks. Python 3.12.3, Pillow/DejaVu, librsvg 2.58.0, pdftotext/pdffonts 24.02.0, pdfinfo 26.05.0. Default/negative audits offline and read-only; reproduction temporary only. Portable relative published commands/no private paths checked.

## Independent second-agent text/routes desk review

A separate agent reviewed the text/routes read-only and found no blocking K02 evidence/privacy/cultural contradiction. It identified three refinements: fresh compound-prompt pacing was untested; ordinary pair kit could confuse individual first-response evidence; timetable fit was asserted without local trial. Fixed by pre-made neutral/oral routes, short prompt parts and unfinished/not-yet-observed rule; explicit separate per-child capture/no peer-answer credit; school-selected/adapted timetable wording. Completion speed remains untested. Incomplete print/technical work was excluded from that independent review; final render/technical checks above are author's executed checks. This is independent agent desk scrutiny, not an educator/child/parent/access/cultural/funder validation or actual K02 result.

**Remaining human/local gates:** educator curriculum/timetable/pacing/language/authentic-event/source approval; genuine element-specific K02 evidence and private family supported/open breadth; current child/family/source/owner/audience/marker/return/privacy permissions; recognised cultural-authority/ICIP where relevant; individual print/screen-reader/tactile/AAC/language access; independent teacher/child/parent/funder evaluation. No actual marking, family knowledge, care/use benefit, viewer learning or completed K02 outcome occurred in authoring.