argument-selection.md

back to table · edit · history · where entries came from · files · download

2930 bytes, as of the version from 2026-11-15 20:42 (current). Recorded here, not run.

# Argument selection for the quadratic central coefficients

The value is gamma_D(m)=m! Lambda_D^(2m)(1/2)/(2m)!, where
Lambda_D(s)=(|D|/pi)^(s/2) Gamma((s+a_D)/2)L(s,chi_D),
chi_D(r)=(D/r), a_D=0 for D>0 and 1 for D<0. Parameter order: D,m.
The natural domain is every nontrivial fundamental discriminant D and every
integer m>=0. D is a discriminant, not a radicand or Conrey index. The
expansion uses z, not i*z, and no s(s-1) factor or central-value division.

- **D: arithmetic_structure.** Goldmakher, Theorem 1.1 and the following
  remarks (PDF page 2), identify precisely the primitive quadratic characters.
  Include every nontrivial fundamental D with |D|<=40, sorted by |D| with
  positive first on ties: -3,-4,5,-7,8,-8,-11,12,13,-15,17,-19,-20,21,
  -23,24,-24,28,29,-31,33,-35,37,-39,40,-40. This covers both parities,
  odd/even and prime/composite conductors, rather than just exceptional
  closed forms. The bound 40 is a noncanonical practical reference cutoff:
  it preserves the family's common comparison range and permits full
  independent checking of small finite character sums at high precision.
  No successful-numerics filter is applied.
- **m: discrete_index.** Include every integer 0<=m<=20, hence even derivative
  orders 0 through 40. A consecutive prefix supports coefficient transforms
  and future Jensen experiments without missing orders. The noncanonical
  cutoff is a moderate initial depth at which independent quadrature and
  cancellation checks remain inexpensive. It is not a count target or a
  theorem about asymptotics. Breadth over all small conductors precedes
  greater depth. Odd central derivatives vanish by the functional equation
  and are stated as a formula rather than repeated zero entries.

There is no free continuous coupling, evaluation mesh, root index or additional
quantity selector. The arithmetic bound and index prefix do not invoke a
reference_sample exemption. This is one quantity at one intrinsic expansion
point. Programs, generator enumeration and completeness metadata must all
specify these same 26*21=546 identities. No positivity, hyperbolicity or GRH
claim is inferred from them.

Checked source conventions: Zagier, Utrecht lectures, Appendix A3, printed
page 60/PDF page 62, gives this exact completion and the quadratic functional
equation. Kedlaya 6.1-6.2 gives the parity-dependent theta Mellin integrals.
Griffin-Ono-Rolen-Zagier, arXiv:1902.07321v2, equation (1), illustrates the
exponential even-series indexing for the pole-removed zeta object; the present
nonprincipal entire Dirichlet object has no pole-removing factor. This is a
convention analogy, not a transfer of their zeta hyperbolicity theorem.

Instruction archive: ../20261115T202357Z-build-skills/manifest.json;
local/live bundle d6661c2f21613524889fa3a57d3c95f8330344ca62e5a856211c08a282b7dbab.
This is the builder's justification; an independent argument review is pending.