# Desk run-through and release limits · 29 September 2026

## Teacher desk trial

I followed Day 3 with only the [B score](print/file-b.pdf), eight scrap cells and the teacher script. The first beat occupies two cells, the full-beat rest occupies two, D4 and E4 each occupy one, and G4 occupies two. That fills eight without a fifth beat. The moved-rest model can be laid down in seven minutes by a teacher who has printed cards; a learner who needs longer should receive it. The script gives a neutral prompt, an independent choice, a count audit and a collectable exit. Day 7 requires two aligned rows; the [F sheet](print/file-f.pdf) and exact text show that the voices overlap rather than add sequentially. The practical risk is treating the optional sine WAV as a performance model. It is only a timing/pitch demonstration; the teacher should supply varied, properly licensed repertoire in the full unit.

## Learner desk trial

I used the [learner desk](LEARNER.md) without the worked key. File A can be read from the diagram or exact text, with no audio. For fresh G, the student page spells out the unfamiliar half-beat gap; a learner must still choose and justify their own variation. For fresh H, the two bars are separated clearly and the directions require counting each one. The six familiar A–F patterns do not expose the G/H event sequences; the day-5/day-10 supplementary swaps now use older practice files and are marked post-check. Public answer keys remain discoverable, so these are formative tasks only. Small cells and long text could still be hard for some readers; large print, raised blocks, a neutral reader and extra time are valid access choices. Silent score reading can show design reasoning but cannot prove live aural perception.

## Potential funder/public-interest review

The source chain is explicit: official QCAA 2025 v1.3 is linked, the exact Unit 1 name and area are stated, event data generates the sheet and WAV assets, and the verifier checks sums, media properties and hashes. I rendered and visually inspected one-voice, two-voice and fresh two-bar sheets. WAV duration and first-note frequency were checked in code; an independent spectral spot-check of F found peaks near its simultaneous E4 and C3. There has been no human listening panel. All bundled melodies and tones are newly authored; no external recording is redistributed. The free CC BY 4.0 notice, font record and rights boundary are visible. This is a small opening starter, **not evidence of improved learning outcomes**. No students, teachers or funders have piloted or endorsed it, and accessibility has been desk-reviewed rather than tested with learners using assistive technology. No translation or First Nations music is included or claimed culturally reviewed. Those require authorised expertise, community guidance and local testing before publication.

## Release decision and next checks

The pack is suitable as a **reviewable public starter**, with the public formative and asset limitations stated on the front page. Before claiming classroom impact, run a teacher pilot with timing/engagement notes and ask learners using different access routes to critique the notation. Before calling the full QCAA Unit 1 covered, add varied repertoire across contexts/styles/genres, documented rights, conventional notation where suited, extended performance/composition/musicology work and school-designed assessment. Re-run `python3 verify_pack.py` after any change; regenerate and rehash when changing an event.