# Teacher, learner, access and funder desk run-through · 29 September 2026

This is a **simulated review of the authored pack**, not a classroom pilot, representative learner study, accessibility certification or funder endorsement. All reviewed student responses below are imagined examples used to expose flaws. The reproducible source/asset checks are in [verify_pack.py](verify_pack.py).

| Lens | Route actually walked | Finding and correction |
| --- | --- | --- |
| Teacher | Opened [Day 11 lesson](LESSONS.md#week-3-day-11-two-rails-one-claim-with-a-hidden-leap), Source B and the [two-rail mat](print/two-rail-claim.pdf). Worked `n=4` on two labelled strips: 96 cm per rail, 192 cm combined. Then followed [Day 17](LESSONS.md#week-4-day-17-a-price-can-change-a-resource-choice): p4 F20/G16, p8 F28/G32, equality p6=24, budget30. | The scripts now demand **per rail versus combined** and **panels per rail n versus total panels p** labels before a recommendation. Quoted-cost conclusions remain conditional on invented prices and unknown quality. A teacher can set up the day from the source, timed steps, route and key without personal data. |
| Fictional Year 8 learner | Followed [Day 13 route C](LEARNER.md#day-13) with the constructed fallback table, then [Day 14 route B](LEARNER.md#day-14) with word strips. Compared one possible reader reaction with actual reported words. | The fallback is labelled constructed on the source, learner page and A4 test card, so it cannot be mistaken for the learner's experiment. The visual task makes two original layouts and a revision; it does not ask the learner to claim one reader represents everyone. |
| Access support planner | Read [all three A4 PDFs](print/TEXT-ALTERNATIVES.md), checked selectable PDF text, then performed the same steps through [full text/tactile equivalents](print/TEXT-ALTERNATIVES.md). Switched the Day 19 response between spoken, AAC and written routes. | The sheets use large labels and non-colour-only distinctions. Full number, order, blank-field and source-label equivalents exist. An exact adult reading is recorded as listening comprehension; a paper wireframe or AAC voice is not falsely marked as browser use or vocal control. Actual device/user validation remains local work. |
| Digital reader | Ran the [offline tool](interactive/model-audit.html) in Chrome from the local file, testing default n4, d3/d4, p8, negative and over-limit entries; observed zero external resource requests. Followed the paper-route link and inspected 1024 px and 375 px layouts. | The numeric outputs and scope notes matched Sources B/D/E. Browser review found a Markdown-style link printed literally inside HTML; it was corrected to a working HTML anchor. At 375 px, the long closed-selector label was clipped, so option labels were shortened while full source details remained in results. There was no horizontal overflow. The tool accepts only 0–1000 whole counts and stores no entries. Keyboard-only navigation and screen-reader behaviour need a real user's validation before any compatibility claim. |
| Prospective funder | Traced the [official 27-row crosswalk](CURRICULUM-CROSSWALK.md), [rights ledger](SOURCES-AND-RIGHTS.md), [fresh check/key separation](ASSESSMENT.md) and [hash manifest](MANIFEST.sha256). Checked whether the practical and economic examples could support a real-world impact claim. | Every curriculum row retains its exact Year 8/Years 7 and 8 level and a partial/conditional boundary. Paper test results, supplier tokens and fictional audience notes cannot prove safety, market behaviour, access or learning outcomes. Broader literature, actual jurisdiction review and community authority remain explicit gaps. |

## Visual print inspection

Rendered all three one-page A4 PDFs to 1200-pixel images and inspected the full pages: no overlapping field, clipped prompt or colour-dependent meaning was observed. PDF text extraction retained labels, calculations, blank-field prompts and licence footer. The paper-test card intentionally leaves all trial cells blank; no held-out check result is printed. Response lines may need enlargement or continued writing space for an individual learner. Font rights are [separate](print/FONT-RIGHTS.md).

## Decision boundary

A passing [verifier](verify_pack.py) establishes dated internal structure, exact pinned workbook rows, arithmetic, fresh-check separation, deterministic SVG/PDF generation, local link integrity and hashed payload files. It does **not** prove student benefit, full Year 8 coverage, an accessible experience for every person, a safe installation, real market data or an approved local syllabus. A school must run its own lessons, observe actual work, moderate it and adapt under local requirements.

**Original review rights:** © NeuroForgeIO Pty Ltd 2026, SubjectNest, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Credit author, source, licence and changes.