# Decision Record ## qmd Step 0 - `timeout 30 qmd query "barrel organ hand-cranked pinned barrel pallets flue reed organ pipes bellows windchest"` started, expanded search variants, searched, began reranking, and timed out without usable returned content. - Result: this packet is grounded in the user request and real barrel-organ mechanics, while keeping all dimensions, tuning, wind, barrel-program, and pipe claims at concept or pending-measurement authority. ## L2 Uplift Decision Promote the packet from L1 concept to L2 V5 build-packet candidate because the repo now has a subsystem plan, expanded estimated BOM, and cut-list scaffold for review. This is not a readiness promotion to build-ready status. All physical specifications remain `pending_measurement`. ## Family Placement The repo remains at `instruments/woodwind/barrel-organ` because the sounding elements are organ pipes: air-column flue pipes and/or reed pipes. The instrument also has a strong mechanical-program component, but current workspace taxonomy places pipe-organ aerophones under woodwind-style folders. ## Key Decisions - Treat the crank as driving both musical program advance and wind production. - Use a pinned barrel or equivalent raised-feature carrier as tune storage. - Keep fixed barrel, removable barrel, and replaceable strip options open until a mechanism mule clarifies indexing and wear. - Use followers or trackers from the barrel to open windchest pallets. - Separate action-mule, wind-mule, and pipe-mule evidence so failures can be diagnosed without guessing. - Keep both flue and reed pipes as possible voice families until wind demand, package size, and voicing complexity are reviewed. - Make the first prototype a mechanism and wind mule, not a tuned musical instrument. ## Open Questions - Should the first barrel be fixed, removable, or represented by a replaceable pinned strip? - How should crank motion divide between barrel rotation and bellows pumping? - What follower material and spring return avoid noisy or sticky tracking? - How will start-of-tune indexing reset without damaging followers or pins? - Is a reservoir needed for steady speech, or can a simple bellows satisfy the first demo? - Should the first musical rank use flue pipes, reed pipes, or one of each as a comparison? - What service openings are required before the case becomes a finished enclosure? - What evidence should promote this packet to L3: reviewed action drawings, measured wind stability, pipe response, or all three? ## L2 Acceptance Checks - README uses an L2 status while preserving non-build-ready language. - `design.md` names subsystem gates without adding dimensions or tuning values. - `bom.csv` and `cut-list.csv` are complete enough for review but keep every physical value `pending_measurement`. - `visual-output-register.csv` continues to classify artifacts only as `concept_only` or `pending_measurement`. ## Authority Boundary No dimensions, pin spacing, barrel scale, pipe lengths, tuning values, wind-pressure values, CAD geometry, DXF coordinates, or measured performance claims are made in this repo. Promotion requires measured action behavior, wind stability, pallet sealing, pipe response, and reviewed design tables.