Skip to content

Tier-1 test evidence: the VAL-05 harness and its closed forms (WP45 item 24)

Tier-1 test evidence: the VAL-05 harness and its closed forms (WP45 item 24)

  • [tested: tests/tier1/test_val05_geometry.py::test_val05_a_uniform_scale_gives_two_s_and_the_exact_ratio] A uniform radial scale r → (1 + s) r maps each innermost crossing to (1 + s) times itself, so Δ/r = s on every plane, the first-order sum ε_G = 2s is exact, and the exact series ratio is (1 + s)² − 1. [tested: tests/tier1/test_val05_geometry.py::test_val05_the_conductance_tolerance_fires_naming_leg_value_threshold_and_z] A 6 % wider lumen is therefore ε_G = 12 %, over the 10 % of 2WCD and the 5 % of the ensemble. Source: tests/tier1/test_val05_geometry.py.
  • [tested: tests/tier1/test_val05_geometry.py::test_val05_the_constriction_tolerance_fires_naming_both_minima] A 0.2 nm notch of four planes moves r_c by 0.15 nm and barely moves ε_G or the rms: the constriction tolerance is the measure that sees a local notch. Source: tests/tier1/test_val05_geometry.py (test_val05_the_constriction_tolerance_fires_naming_both_minima).
  • [tested: tests/tier1/test_val05_geometry.py::test_val05_no_plane_sits_on_a_reference_tip] The reference's z extent of 29.5 h rounds to 30 planes; the mid-planes sit 0.025 nm off each 0.3 nm vertex (0.3 × (1 − 0.025/0.5)). The crossing test is half-open, so the plane on the top tip is never crossed and is dropped, or the reference would be refused against itself. Source: tests/tier1/test_val05_geometry.py (test_val05_no_plane_sits_on_a_reference_tip).
  • [tested: tests/tier1/test_val05_geometry.py::test_val05_each_constriction_is_located_on_its_own_polygon] The constriction window's Δr_c is the difference of two minima at different z, not a difference at one plane. Source: tests/tier1/test_val05_geometry.py.