back to table · edit · history · where entries came from · files · download
2708 bytes, as of the version from 2026-11-18 22:18 (current). Recorded here, not run.
# Argument selection for T753
Quantity: regular Laurent coefficients at the fixed double pole s=1 of the
secondary zeta function Z(s), using alpha=(rho-1/2)/i, every upper-half-plane
nontrivial zero with multiplicity, principal powers, and meromorphic
continuation. There is no RH or simplicity assumption.
Ordered parameters are form (Symbolic), then n (Z). The natural domain is
form=coefficient or derivative and every integer n>=0. For the regular part
F(z)=Z(1+z)-1/(2*pi*z^2)+log(2*pi)/(2*pi*z), these select c_n=[z^n]F(z)
and C_n=F^(n)(0)=n!*c_n, respectively. These are invertible conventions of one
number at the same index. Both are compact through order 20; the factorial
conversion is useful for identifying a value recorded in either convention.
The duplicate forms at n=0 and n=1 are intentional.
Both parameter assessments are discrete_index. The form parameter is a finite
normalization index, not a choice of different functions; n is the actual
nonnegative Taylor/derivative order. Every n=0,...,20 is included, coefficient
before derivative, for 42 entries representing 21 distinct coefficient orders.
The finite cutoff 20 is a practical choice limiting Cauchy extraction and
independent high-order cancellation cost. It retains the finite part and
twenty consecutive nonconstant terms, with both conventions at every order.
It is not a mathematical domain restriction or a quota. No useful independent
continuous parameter family is omitted: the pole and normalization are fixed.
There is no grid, shift, quantile, coupling, or exploratory reference_sample.
Checked sources: Kawalec, arXiv:2403.15741v2 section 4, equations (64), (65),
(69), (71), Table 9; Arias de Reyna, arXiv:2006.04869v1 introduction (1),
for the unconditional alpha definition and pole locations; Voros,
arXiv:1403.4558v2 equations (13)--(19) and (25), for the confluent family and
central completed-product conventions. Kawalec's real-ordinate/RH wording is
not adopted as an unconditional equality with the separate ordinate harmonic
constant. The shared proposal #342 and dossier are research leads, not proof.
The generator, Programs, flat stored row order and complete-note must all keep
this exact selection. The initial draft's complete:no remains correct because
all n>20 are absent. The independent critic must check these assertions and
bind its own sidecar to the current live selection and all generator sources.
Skill evidence: ../20261118T135803Z-build-skills/manifest.json, bundle SHA-256
d6e85d49873fac5bc748d068b0288cf242b0f9aebf3a9a0fff70295019227420;
local and live resource hashes were validated and match. skill-manifest.json
and skill-verification.json preserve the record with this run.