Artian's HVP Source-Access Reference Framework

Which HVP object is the source, which one is the apparatus, and which one is only statistics? The paper-native BaBar source-access word is frozen before the public row is judged: \[ F_{\rm BaBar}^{\rm QTT}(s)=e^{-4\chi_0}\left(\frac{R_{2\pi}^{\rm src,QTT}(s)}{G_{\rm src,BaBar}}\right)^{-4\chi_0},\qquad \chi_0=\frac{15}{512\rho}. \] Version 52.0 types the complete route from source to access word, calibrated physical camera, finite binning, nuisance quotient, and covariance verdict. A camera attenuation is not silently relabelled as the canonical Access-Law projector. Two results are simultaneously true. In the primary BaBar window, the integrated scalar is \[ a_{\mu,2\pi}^{\rm QTT}=371.643570327919\times10^{-10},\qquad z_{a_\mu}=-0.593757, \] a scoped QTT zero-target-fit empirical victory for the scalar direction. The same frozen source fails the resolved lineshape: \[ \chi^2_{\rm sh}=85030.2610802301,\qquad \nu_{\rm sh}=139,\qquad \chi^2_{0.95}(139)=167.514304991. \] This is a frozen-source failure, not a quantity available for post-hoc source repair. The scalar result and functional failure are different covariance observables and neither is allowed to erase the other. Verification: 28/28 typed-access checks and the inherited 31/31 independent HVP checks pass. The full 360-page execution history is retained.