CONSTRUCTION ALTERNATIVES AND MEASUREMENT DECISION BOUNDS

This retrospective addition accompanies S9.9 and main section 3.1, v1.54.
It leaves all previously reported scientific results unchanged.

Inputs are the unchanged main Tables 1 and 2. The original operation library
contains 23 explicit lengths from the 10:24:26 right triangle. Every ordered
assignment of four distinct entries is evaluated, with scale fitted solely
to the exterior. fixed_assignments.csv.xz stores all 212,520 scores and indices;
results.json defines the index-to-operation map and reports the comparisons.

The systematic extension adds every positive pairwise difference between
sides, half-sides and quarter-sides. Equal numerical lengths are collapsed,
while all expression aliases are retained. The resulting 42 distinct lengths
give 2,686,320 assignments; the same four predicted lengths are first. A
componentwise nearest-length proof and independent brute-force verifier agree.
This ranks dimensional predictions, not historically unique procedures.

Free scale is reported with two explicitly different estimators:
1. minimize relative RMS itself (target rank 7);
2. fit in metres by absolute least squares, then score relative RMS (rank 6).
No free-scale number is substituted for the exterior-calibrated comparison.

Decision cells show the observed measurement at which another predicted
length becomes closer. They produce exact open-box tolerance bounds in the
stated finite libraries. Sensitivity boxes are stipulated scenarios, not
measured survey uncertainties or confidence intervals. The coupled exterior
bounds retain the fixed ideal frame; positive least-squares coefficients
propagate bounded relative exterior perturbations into the same scale bound.

Reproduce with Python 3, NumPy and SciPy:
  python analyze.py
  python verify.py

The enumeration was specified after the selected construction and measurements
were known. It tests alternative operations and assignments conditional on
the stated inputs and families, without correcting the entire research search
history or establishing prehistoric intent. No original astronomical census,
Woodhenge target, measurement or historical file is changed by this addition.
