# Teacher, learner, assessor and funder desk run · Science Weeks 27–28

**Teacher desk run:** I can print K and the three swatches, read K before L, and teach ten 25-minute sessions without pretending a physical sample is present. If the school approves the exact broad card/fabric/corrugated-card materials, I verify source labels, pretest access and fibre/edge/adhesive safety, then log local date, raw visible feature, direct access or attributed ADULT REPORT and unknown joins. If not, **REAL OBSERVATION NOT AVAILABLE** is the correct record and the full FICTIONAL MODEL route remains. I keep the Day 135/140 checks unseen until independent first responses. I do not use the integrated maths balance to infer materials or a HASS preference interview to infer texture.

**Learner desk run:** In K I can identify the plain-looking half and six lines, but I cannot name its real materials. In L's made-up note, A is reported card, B fabric and C paper-based card; B and C are both reported ribbed, so “ribbed means fabric” fails there. I may hear, see, use AAC/sign/voice, point, direct an adult or choose a secured broad tactile *paper* diagram. I do not have to touch fibre, speak publicly or bring anything from home. The maker can say “I don't know”; then I keep that part unknown. Family options cost nothing and do not require returning.

**Assessor desk run:** AD's TRIANGLE PANEL 64/four dots/three zigzags and AE's PATCH NOTE 91/P–Q–R dot set are absent from instruction, learner routes, worked swaps and public print aids before first response. I log exact first words, access and neutral reread before seeing the separate key. A picture region does not prove material composition; a maker's fictional label remains attributed; a tactile paper dot does not become a physical fabric feature. Actual AC9SFI02 credit requires separate approved dated genuine inspection. One or two paper checks do not establish term/year mastery, child growth or a real property.

**Potential funder desk run:** The pack supplies ten distinct scripts, thirty response routes, twenty worked examples, ten optional family prompts, three narration-ready text masters, two fresh checks with keys, six illustrated A4 SVG/PDF pairs, exact ordered text/tactile routes, source/rights boundaries and a regenerable hash receipt. Those delivery facts are inspectable. Learning gains, teacher time saved, certified accessibility and state-by-state endorsement are **unmeasured** here. Human educator/accessibility review and consented classroom pilot with predeclared measures are next.

**Visual print critique (30 September 2026):** rendered and inspected all six A4 PDFs at reading size. The split regions, different B/C stroke patterns, schematic safe table, blank actual-source fields and two speech bubbles are pictorial rather than label-only. The first sort mat placed empty sample squares across the bin divider and could be mistaken for pre-sorting; they were moved into a separate UNSORTED tray, then the print was regenerated and re-inspected. [Verifier](verify_pack.py) checks geometry, core text, held-source separation and hashes; it cannot certify screen-reader behaviour, braille quality, child comprehension or classroom effectiveness.

**Next human checks:** local material/product provenance and allergy practice; neurodivergent and disability access with affected learners; full-scale classroom print; check prompts for cueing; applicable state syllabus/assessment; consented classroom use. These have not been represented as completed.