argument-selection.md

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

3398 bytes, as of the version from 2026-11-16 20:19 (current). Recorded here, not run.

# Argument selection for T726

Quantity: Simonič's $\widehat g_n$, Theorem 1, equations (2) and the adjacent definition, arXiv:1612.08627v2, p. 2. The pair consists of consecutive positive simple zeros $\gamma_n,\gamma_{n+1}$ of Hardy Z, counted from one with multiplicity. All derivatives are with respect to real height. The analytic log-gamma phase is real at 1/4 and theta(0)=0. There is one free parameter, `n: Z`, with n >= 1 and both endpoints simple. No second endpoint, derivative order, coupling, probability, or mesh is selected independently.

The natural family is infinite. The computed selection is exactly 1 through 100, followed by 6707 through 6711, in increasing order (105 entries). Every pair is retained regardless of the resulting condition. The 100-gap prefix is a practical, noncanonical cost bound for broad consecutive low-height reference coverage, including ordinary gaps and failed sufficient conditions. It is not a mathematical endpoint or quota. Moving to higher zero indices is meaningful depth within this one discrete family; introducing arbitrary gap combinations or other functions would not improve the family's breadth.

The parameter has mixed selection roles, partitioned as follows:

* `documented_application`: n=34 and 6709. Simonič's Tables 1 and 2 (p. 7) explicitly select the first and historical Lehmer pairs. These checked source examples motivate the normalization and difficult cancellation controls.
* `discrete_index`: every integer n=1,...,100 except 34. The adjacent pair's lower ordinal is a natural discrete index, and this prefix provides comparisons without selection by the answer. The finite bound is computational.
* `reference_sample`: n=6707,6708,6710,6711. Exactly the two neighboring gaps on each side of the historical gap. The purpose is exploratory recognition and local comparison of an exceptionally close pair with its immediate environment. This bounded contiguous window avoids an expensive second high-height prefix. Its selection is noncanonical and independent of decimal spelling. The paper defines the diagnostic for arbitrary simple pairs and names the central pair; it does not claim special applications for the four neighbors. Source definitions and the historical-pair location were inspected directly.

The reference policy is: purpose = local recognition and comparison around the historical gap; bounds = exactly 6707,6708,6710,6711; coverage rationale = two adjacent gaps on each side, selected before evaluation, without answer-based pruning. Noncanonical acknowledged, notation independent, sources checked. Explicit memberships are in selection-groups.json; that file is builder rationale, not an independent verdict.

`complete: no`. Enumeration, Programs, stored identities and complete-note must agree on the selection above. The finite counting certificate and nonvanishing derivative enclosures establish numbering, consecutiveness and simplicity for selected inputs without assuming RH. The source's implication to Lehmer pairs remains conditional on RH; the numerical definition does not need global RH or simplicity.

The other two proposed conditions in family #324 are distinct mathematical quantities and are not stored here. T584 stores their raw Hardy derivative ingredients, T3 the zero ordinates, T585 stationary points, and T586 signed values at landmarks. No published table is changed.