nx_unitconv.nx

nx_unitconv.nx -- MEASUREMENT SYSTEMS + EXACT CONVERSION (engineering-twin metrology substrate). (Placement note: DISTINCT from nx_units.nx = kernel_v2 SI dimensional-ANALYSIS engine (physics equations, 7-D dim vectors) and nx_measure.nx = Lebesgue measure theory. THIS organ converts between the world's still-used measurement SYSTEMS -- metric, US customary, UK imperial, survey -- on the runtime/i64 stack. They compose, not compete: analysis checks equations, THIS checks the numbers feeding them.) The anti-Mars-Climate-Orbiter organ: that probe was lost because one team produced lbf-s and the other READ it as N-s (factor 4.4482) -- a silent unit mislabel. Doctrine: (1) every value travels WITH a unit index, (2) conversions are EXACT RATIONALS from the international definitions (1 in = 25.4 mm EXACT since 1959, 1 lb = 453.59237 g EXACT), never floats, (3) every conversion returns an EXACTNESS REMAINDER (0 = exact) so rounding is VISIBLE, (4) dimension mismatches and overflow REFUSE LOUDLY (error codes) -- never a wrong number. Covers metric (SI), US customary, UK imperial (gallon/pint DIFFER from US -- still in daily use), the US-survey-foot trap (1200/3937 m, 2ppm off the international foot -- state-plane data hazard), and the twin's own fx256-mm grid as a first-class unit. Bases integer-exact: LENGTH nm, MASS ug, FORCE nN, VOLUME uL, TIME us, IMPULSE nN-s. Data-driven registry: units are TABLE ROWS, auto-gcd-reduced + overflow-guarded at add (fail-fast). license_tier: ORIGINAL