# Desk run-through · Foundation mathematics Weeks 39–40

**29 September 2026 · authored draft, no teacher, child or class observed.** I worked through the [scripts](LESSONS.md), exact [practice sources](MATERIALS.md), [clean learner routes](LEARNER-CARDS.md), [fresh check prompts](STUDENT-CHECKS.md), separate [public key](teacher/KEY-AND-NEXT.md), and nine generated print pages as a teacher, learner, time-poor supporter and potential funder/reviewer would. This was an offline **desk simulation**, not a physical print/cut trial, classroom pacing study, individual AAC/tactile assessment or evidence of learning gain.

| Role and case | What I actually did | Finding and remaining limit |
| --- | --- | --- |
| Teacher, Week 39 explanation | Traced P191's eighteen source circles to numeral 18 on the complete 0–20 rail, with 17 before/19 after and empty frame representing zero. Rebuilt P192's 3+5=8; moving one from the three-part yields 2+6=8 without altering whole. Followed the P193 fixed SHELF/BOARD mat and P194 given 7 KITE/4 FAN question. | The number, part, shape/place and picture-data **strategies** are distinct; a paper visitor is labelled model audience. Real listener feedback is not invented. The ten scripts are timed on paper only; a real class may need different pacing. |
| Teacher, Week 40 review | Replayed D196's separate 6+4=10, 6+2=8, 8−3=5 actions; D197's full CIRCLE–SQUARE–TRIANGLE unit; D198's same-page 180/250 drawing-unit strips and four day-part labels; and D199's five groups of two plus separate 5 PEN/7 BRUSH picture source. | Group size and group count stay separate, as do operation sources and data sources. Direct printed length has a same-scale boundary; capacity, mass and measured duration remain **not assessed**. No false claim about real stock, routes, seasons or child routines is made. |
| Learner, routes A/B/C | Followed all three D198 modes on the same printed source: align/overlay M and N, trace both from one common start or direct an adult to align; then order MORNING, LUNCHTIME, AFTERNOON, NIGHT. All routes reach “N extends farther on this page” with an endpoint reason and the same time sequence. | No speech, fine-motor cutting, hearing, colour discrimination or personal day disclosure is required. Actual individual access and physical print scale require local checking. |
| Learner, fresh Check A before the key | Used only the child prompt and Check A page. Counted two rows of eight round source marks, checked the separate 17 claim, named the circle's curved boundary and its relation above fixed BOARD, then sorted the fourteen pictograms. Afterwards compared to the key's 16, 9 PINE/5 SHELL and four unmatched PINE results. | The child-facing source does not print corrected totals or a key link. Three actions are feasible as compact samples, but a real child may need the teacher to spread work across time. The source and key are publicly reachable, so this is formative, not secure. |
| Learner, fresh Check B before the key | Used the Check B page's 2/6 PART A/B squares, 7 START/3 ADD circles, two printed SQUARE–CIRCLE–CIRCLE units plus three blanks, and common-start A/B strips. Confirmed with the key: whole eight, end ten (repair display nine), whole next three-card unit and B longer on the same page. | These cases are distinct from practice and English-held text sources. It samples four targets, not the complete Foundation achievement standard. The right result alone is weaker evidence than a visible object/move trail. |
| Time-poor supporter | Tried no-cost drawn variants: an eleven-mark count/rail explanation and a seven-mark part split; checked the child card can be used privately without a printer, internet, purchase or real-person survey. | A supportive adult can read and move pieces under child direction. No home participation or engagement was observed. |
| Potential funder / rights reviewer | Traced twelve exact Foundation maths descriptions through dated [official source snapshot](source-snapshot.json) and [partial crosswalk](CURRICULUM-CROSSWALK.md), checked the [QCAA Prep v9 entry point](https://www.qcaa.qld.edu.au/p-10/aciq/version-9/learning-areas/p-10-mathematics) and separated official wording from original source/data/vector rights in the [ledger](SOURCE-AND-RIGHTS.md). | This is a reviewable two-week year-end sample, not a full-year quality/effectiveness proof, school achievement judgment, state approval, translation review, First Nations resource, accessibility certification or secure exam. |

## Print, arithmetic and automated evidence

- Generated **nine original single-page A4 SVG/PDF aids** from [local vector code](print/generate_print.py). Rendered every PDF to PNG and visually inspected a nine-page contact sheet. The numeral rail, broad mat frames, fixed landmarks, check source marks, pictograms, strip starts/endpoints and footer remained visible without clipping. PDF text is selectable but **untagged**; [full text/tactile production routes](print/TEXT-ALTERNATIVES.md) are the reliable alternative. No physical print or assistive-technology trial occurred.
- Independently recounted Check A's 16 marks, 9 PINE/5 SHELL and fixed circle/BOARD relation; Check B's 2+6=8, 7+3=10, two complete three-shape pattern units and 145/230 drawing-unit strips. The [verifier](verify_pack.py) counts machine-readable SVG attributes and checks A4 PDF metadata/selectable text; only actual same-scale physical output can verify a local printed length comparison.
- Ran broad Ruff `E,F,I,UP,EXE,PLR,RUF` over all new-pack Python, Ruff format check and the fail-closed [verifier](verify_pack.py). The verifier checks the official workbook hash, twelve exact code/row/wording matches, 10×25-minute scripts, 30 choice routes, 20 worked swaps, two separate public checks, nine aids, local links and the [SHA-256 manifest](manifest.json). These gates prove internal artifact consistency only.

## Local educator handover

Check current Queensland Prep/school sequencing and actual achievement-standard expectations; print the length sheet at a consistent scale; inspect broad pieces for the particular child; try real AAC/tactile/directing routes; collect first child-controlled responses privately; mark adult mathematical cues, missing samples and unassessed strands honestly. Any real listener response should be recorded only when it occurred. Do not upload authentic child work to the public library or replace year-long evidence with these two checks.