nx_sim_validation_gate.nx

nx_sim_validation_gate.nx -- climb the VALIDATION axis with the ASME V&V 20 validation-metric METHODOLOGY: comparison error E = |S - D|, validation uncertainty u_val = RSS(u_num, u_D), and the validated-within- uncertainty test |E| <= k*u_val. Here S = our fixed-point LEAPFROG oscillator at time T; D = an INDEPENDENT referent (the exact analytic x(t)=A cos t via our sovereign cos -- a different method than the integrator); u_num = Richardson discretization-error estimate |x(dt) - x(dt/2)|; u_D = the referent's own error bound. HONEST SCOPE (a gate row): the referent is an exact ANALYTIC solution, NOT a physical experiment, so this demonstrates the V&V 20 validation METHODOLOGY against an independent referent -- it advances Validation to BASIC, but full NASA-L4 / regulatory credibility still requires REAL-SYSTEM experimental data (the L4 ceiling). LIAR-KILL: a wrong-physics model (wrong stiffness k=1.2) gives |E| >> u_val and is REJECTED. GREEN iff 6/6. license_tier: ORIGINAL