Exploratory extension, 2026-09-29. Not part of the original frozen protocol.

Question: Does the previously specified D=g class retain its position when
every original R10.1 grammar expression uses h=K/L, K=2*li*wi*di/We,
as in the already reported R13 joint reconstruction?

Keep all 1950 original expressions, segment labels, and the physical domain
30<L<W-30. Compare to L0=90.60 in and h0=41.31 in with the existing 0.2%
tolerance. Report length-only and joint acceptance separately; h=K/L makes
these output residuals dependent. Neither count supplies a probability.
Keep every real admissible root. Multiply by the common positive denominator
2*L*(W+L), giving polynomials of degree at most 3; use exact rational
coefficients and numerical polynomial roots. Exclude exact tautologies and
constant residuals. Check the original fixed-height panel against its saved
127 classes before interpreting the extension. Use original class grouping
by rounded roots, and also report minimum separation between classes.

Vector question: Calculate the unit vector of the actual 2D X branch and
compare to the reported profile inclination and the ideal 11:14 slope.
No scanning of coordinate conventions, star pairs or celestial epochs.
Any equality derived from D=g is algebraic structure, not another observation.
Absolute celestial direction requires measured azimuth, reference frame,
epoch and celestial target not selected by this calculation.

Report unfavorable as well as favorable outcomes. Do not change the
submission package automatically.
