nx_myoattach.nx

nx_myoattach.nx -- ★MUSCLE ATTACHMENTS DERIVED FROM BONE LANDMARKS, never typed. ★THE LAW THIS APPLIES, learned in this lane on the brow ridge: "NEVER TYPE A PLACEMENT COORDINATE THAT THE PARENT ALREADY KNOWS." A muscle does not have a height; it has ATTACHMENTS -- an origin on one bone and an insertion on another -- and its height is a CONSEQUENCE of where those bones are. nx_myotable still carries typed per-mille stations, which means every future muscle is another opportunity to type a plausible-looking number. This session already measured what that costs: the biacromial breadth sat at 52% of its true value for the life of the organ because it was typed. ★★THE MECHANISM THAT MAKES ~700 MUSCLES REACHABLE WITHOUT 700 GUESSES: an attachment is (joint_a, joint_b, t) -- a point t per-mille of the way along the bone spanning two joints that nx_skelgen ALREADY EMITS. So an attachment is specified ANATOMICALLY ("the proximal third of the humerus") and its coordinate is derived, at any stature, any sex, scale-free, forever. Typing stops being possible for the quantity that was being got wrong. ⚠AND IT IS A CROSS-ARTIFACT CHECK, which is why it earns its place now rather than later: the derived station and nx_myoset's typed station are two independent statements about the same muscle. This organ REPORTS the disagreement per muscle. Per the lane's law, a disagreement NAMES A NUMBER TO GO LOOK UP -- it does not tell you which side is wrong -- so the teeth here are on the MECHANISM and the disagreements are emitted as EVIDENCE, never silently "corrected". nx_myoattach report <skel.dat> nx_myoattach selftest <skel.dat> license_tier: ORIGINAL expect_exit: 0 No hw writes (Rule 26).