# Four-role author desk critique · 30 September 2026

**Status:** author simulations of teacher, student, parent/carer and potential funder only. No independent participant/classroom, actual child/place/account/care record, real display try or physical tactile build. Roles trace files and source relationships; they aren't interest, learning, teaching, access, consent or market evidence.

**Teacher role:** I traced the ten scripts on screen: source-backed feature pages; attributed particular-person reason flaps; ask-first/care flap; two planned arrangements and pass-friendly offer; fresh check; one bounded try; separate second account/gap; first/revised instruction; useful care handover; fresh check. The actual workflow gives safe familiar-place approval and view, willing significant-feature account, exact local care procedure, distinct owner/reader permissions, temporary approved page placement/return or private owner-review kit, renewed willing trial/retry, local record and K03 element judgement. The HASS place feature/meaning/care remains attached during layout trials; a design improvement doesn't replace actual place knowledge. Second source/try isn't compulsory. Seven-minute activity needs pre-made bases/folds and adult-directed routes; genuine local timing, curriculum and source suitability remain educator review.

**Student role:** In author desk simulation I could direct flat broad cues, point/scribe or hear all exact words/diagram relations without touch, folding, walking or a reader role. I could make a useful kit even when actual display/try wasn't approved. Edda and Finn gave their own reasons; Jori's point and Mell's question stayed separate reports with gaps. I could keep first words, repair a sourced feature name, ask a willing reader to try/pass, then leave retry blank. Pass/silence didn't supply agreement or a reason. Fresh sources required same source tasks with different features/people/layouts/responses. These are author interpretations, not learner comprehension, enthusiasm or independence. No physical kit/tactile trial occurred.

**Parent/carer role:** I checked ten free optional invented-paper home bridges, school equivalents and no returned score/evidence, private home inspection/address/photos, device/purchase, family interview/cultural disclosure or real property handling. Skipping has no assessment cost. Actual-source work is educator-approved; home responses aren't automatically authentic evidence. Any home language can carry reasoning, but supplied translations need fluent local review. No parent/carer reviewed this pack.

**Potential funder role:** I checked ten timed scripts, thirty same-goal choices, twenty worked real-life domains, ten bridges, two held fresh checks/key, nine original print pairs/exact alternatives, complete actual K03 opportunity/record, exact primary row pins/rights/hashes. These support an authored/reproducible draft, not classroom time saved, demand, full-year coverage, completed knowledge, fairness, universal access, care benefit, learning gain or commercial suitability. Actual local evidence is absent; K01/K02/K04 remain separate source/authority gates.

**Residual critique:** a point to a page doesn't prove understanding; a question doesn't prove failure. Different readers/layouts/conditions aren't a causal comparison. Equal pages aren't fairness/consent/access evidence. First-direction capture and renewal of willingness require adult scribe/attention. Preserve place feature, particular-person significance and local care in every actual kit and handover; avoid drifting into a display-design test. Some K03 elements may be evidenced while others remain open. Untagged PDFs require individual screen-reader/print/tactile/AAC/language/pacing review. No quality score, child outcome or accessibility certification claimed.

**Reproduction:** `python print/generate_print.py --regenerate` deliberately writes only nine named SVG/PDF pairs and exact alternatives. `python verify_pack.py --manifest-json` prints proposed hashes for deliberate author saving; normal verifier is offline/read-only. `python verify_pack.py --negative-checks` rejects 23 in-memory mutations and restores the normal audit. `python print/generate_print.py --check-reproduction` generates in temporary storage and compares 19 artifacts while preserving pack bytes/size/mtime. Python 3/Pillow/DejaVu, librsvg, pdfinfo, pdftotext and pdffonts required. Fixed PDF build clock is reproduction metadata, not actual source/event time.

## Final technical receipt · 30 September 2026

- **Offline audit:** PASS — nine exact direct workbook XML rows/imported framework/crosswalk; current QCAA alignment/sequence pins; ten 25-minute scripts with eight integrated actual-place branches; conditional K03/no actual outcomes; 30 routes; 20 worked variations; 10 bridges; two held fresh checks/separate key; nine A4 pairs/exact alternatives/measured text bounds/embedded searchable fonts/keyed feature/same-page/layout geometry; 55 local links; 37 SHA-256 files (38 total including manifest).
- **Reproduction:** second deliberate regeneration matched nine SVGs/nine PDFs/full alternatives byte for byte, 19/19. The supplied temporary-storage reproduction mode also passed 19/19 and preserved every pack file's bytes, size and modification time. Fixed PDF build clock is metadata only.
- **Read-only negative checks:** 23/23 in-memory mutations rejected: workbook digest; exact official description; K01 hold; blanket K03 hold; invented actual outcome; timing; ordinary actual-branch removal; useful actual placement removal; willing particular-person source removal; compulsory second account; removal of no-actual-evidence limit; silence treated as agreement; missing renewed willingness; fresh speaker leakage; learner-key leakage; exact alternative drift; printed/source quotation drift; QCAA alignment pin; QCAA sequence pin; source square geometry; fresh above/below feature order; same-page dimensions across plans; stale receipt. Supplied `--negative-checks` mode restored the normal audit and preserved all pack bytes, sizes and modification times. No bad fixture written.
- **Print inspection:** all nine PDFs rendered and visually inspected together. Source shapes, two planned layouts, two attributed reasons, care illustrations, separate reported tries and blank logbook/actual fields were visible without clipping or panel overlap. Final adjustments separated the footer symbol/words and gave stacked cards border margins; both plans now depict the same page dimensions. Place/layout/fresh pages were re-rendered and inspected individually. Exact visible SVG/PDF/alternative text, source quotations and measured bounds passed. No physical tactile or learner access trial occurred.
- **Current primary evidence:** official live ACARA workbook HTTP 200, 2,049,917 bytes and pinned SHA matched. Current QCAA Prep alignment direct browser open and P–6 sequence successful result reopen compared all nine Prep descriptions. Sequence repeat direct URL was restricted; no successful direct PDF download claimed. Current indexed official HASS entry checked; full-page fetch not claimed. ACARA terms rely on earlier same-session successful review, not a new W35–36 terms fetch.
- **Code/runtime:** both original Python tools pass format/lint checks. Python 3.12.3, Pillow/DejaVu Sans, librsvg 2.58.0, pdftotext/pdffonts 24.02.0, pdfinfo 26.05.0. Default audit and in-memory negatives are offline/read-only; reproduction writes only temporary outputs. No learner data collected.

**Remaining human/local gates:** educator pacing/language/adoption/timetable and curriculum decisions; actual place/view/source/privacy/material/task/audience/owner/reader approval; genuine element-specific K03 records; individual print/screen-reader/tactile/AAC/language access review; recognised cultural-authority/ICIP permissions for relevant extensions; independent classroom/parent/funder evaluation. No actual place teaching, care action/benefit, display use, learner knowledge or completed K03 evidence occurred in authoring. No classroom/access/cultural/commercial certification implied.