# Author desk run-through · teacher, learner and potential funder

**Mode:** simulated paper/website walkthrough, 29 September 2026. No child, family, teacher outside the author team, or funder participated. This is a defect-finding exercise, not an outcome study.

## Teacher tomorrow

I open the [start route](README.md), check the two original sources and choose Day 13. I have a blank A4 mat, broad cards and the book-swap row: 46 groups of 18. The lesson gives a 5-minute retrieval, 7-minute model, 10-minute child route, 9-minute challenge and 4-minute evidence note. The [key](TEACHER-KEY.md) has 828 cards and 72 below limit with a second product route. I can put only [Day 13 clean card](LEARNER-CARDS.md) in a child stack. The setup needs about two paper piles and a readable Source A; no printer is required because the [text route](print/TEXT-ALTERNATIVES.md) replicates the mat. I must fit 35 minutes *in addition* to English/maths subject blocks and any other required subjects.

## Learner with two participation routes

I choose the games plan because I like invented stations. With paper I write 24×37, benchmark 25×40≈1,000, then calculate 888 and explain why the benchmark overshoots. If writing is slow or painful, I point to factor and place cards and direct an adult to record my partition. My mathematical choice is still mine; adult help is logged. A heard source can show what I understood, but it cannot be called independent reading. The flyer cannot tell me who prefers anything. I may decline public presentation, use a private visitor, or make a text/AAC recommendation. The [learner cards](LEARNER-CARDS.md) have no 828/888/928 answers or key link.

## Funder or reviewer audit

I can follow the public [exact-source crosswalk](CURRICULUM-CROSSWALK.md) to ACARA row and wording, reproduce the [hash check](verify_pack.py), read the ownership and font ledger, and see both the lesson count and the **unbuilt remainder** in the parent year. The 900-card model deliberately avoids unverified cost, popularity and local-place claims. The measurable deliverable is ten authored project blocks with original source and print aids; no classroom effect, state compliance, inclusive-access certification or full-year coverage is claimed. The teacher key is publicly inspectable but labelled staff copy; schools must withhold it until using an assessment. Here Days 15/20 studio examples are explicitly guided after the subject checks.

## Defects caught and corrected in this pass

- A rough 25×40≈1,000 estimate might be mistaken for proof that 24×37 breaches the cap. Day 12 now says the estimate triggers an exact calculation, and the key shows 888 fits.
- The garden flyer uses “easily fits” although 29×32=928. Day 14 has learners test and fairly correct that claim, while Day 18 preserves the original factors before showing a *new* 28×32=896 plan.
- A polished recommendation could hide no arithmetic or source check. Day 19 requires an inspectable product, rule and source limit; the visitor tries to reproduce the reasoning.
- The print PDF may be readable on paper but lacks a tagged-PDF claim. A complete plain-text/tactile route travels beside it and is the digital fallback.

**Next human gate:** a Year 5 teacher and learners with varied access routes should run an actual block, record timings/first responses and revise. This run-through does not substitute for them. Original SubjectNest review prose © NeuroForgeIO Pty Ltd 2026, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); credit, link and note changes.