By @josebrox
# Physical detector coordinates and quadratic compilation
The [scaling theorem](../negative_results/papers/quartic/physical_detector_scaling_20261117_v67.tex) proves two obstructions for specified quadratic presentations of the revised general-quartic factors. The [construction records](../data/results_20261117_v65.json) remain unchanged: 25 general variables, 11 under ordinary abc, quartic counts 195 and 68 respectively, and 229 for the fully specified quartic. Every count includes the input.
## Physical normalization
In the published 195-coordinate table, put `u=x77` and `v=x90`. The two physical contact weights and the retained unit helper are
```math
q_0=u,\qquad q_1=uv,\qquad x_{92}=u^{-1}v^{-2}.
```
The independent coordinate audit composes all 198 source coordinates through the published Laurent, parameter, primitive-basis and affine maps. It verifies 71 literal test, factor, product, trace and pole equations up to exact nonzero Laurent factors. In detector `j=0,1`, the five physical factor outputs have unit direction `(1,2j)` and fixed-parameter powers `(1,1,1,0,-1)`. The corresponding three partial-product powers are `(3,4,3)`. The audit includes the affine Lucas centering and the distinction between full and half imaginary components.
## Two exact obstructions
For a shared special factor, the two physical contact directions force a residual unit-exponent difference `(0,-3)`. This difference is absent from both stated unit bases, excluding that shared allocation.
The stronger obstruction uses one literal rational reference row and three field factors. Projecting a Laurent exponent `(a,b)` to `a+b`, their quadratic terms force
```math
r_{01}-r_{00}=12.
```
The reference row and the first factor imply `r00>=0`. The constant in the second factor implies `r01<=6` for the original unit basis and `r01<=2` for the modified basis. Both contradict the forced difference. Full-square monomials force equal real and imaginary row multipliers, so that convention does not restrict the allowed diagonal transformations.
The proof permits arbitrary real scaling weights and uses no bounded search. Ten nonnegative integer multiples of necessary degree inequalities give the exact contradictions `0<=-12` and `0<=-20`. The selected factor monomials are reconstructed from their physical formulas and cannot cancel against the distinct factor outputs.
These results concern the displayed factor allocations, the literal reference equation, diagonal Laurent coordinate changes and monomial row division. Other factor pairings, a different reference equation basis, row combinations and nonlinear coordinate substitutions remain open. No general lower bound on defining variables is asserted.
## Verification
Run `python -B scripts/verify_v67.py`. The [portable package](../negative_results/computation/physical_detector_scaling_v67) contains the exact support model, integer certificates and coordinate-map checkers. The core support and selected-map checks use the Python standard library. The independent full-coordinate audit uses SymPy 1.14.0 and imports no construction code.
The [metadata](../data/criteria_20261117_v67.json), [integration receipt](../data/audits/integration_20261117_v67.json) and [rendering receipt](../data/audits/tex_build_20261117_v67.json) state the verified scope. The full legacy verification is also retained. Neither the targets 21 and 9 nor a definition of every number-field ring of integers follows from these compilation obstructions.