research.md

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

3142 bytes, as of the version from 2026-11-17 14:09 (current). Recorded here, not run.

# Guillera hypergeometric periods: checked research

Checked DPVZ author's June 2020 PDF, §2 eqs. (2.1)–(2.3), paragraph before
Table 2.10 and Table 2.10 p.5, against the arXiv v3 HTML table with the same
caption. The roster has fifteen rational rows and a quadratic ± row. The lower
arithmetic multiset consists of five ones; the classical function has four
lower ones. The source's question marks denote experimental identities. None
of those identities or the conjectural Asai identification is assumed here.

Checked DLMF 16.5.2: successive Euler integration gives the real continuation;
choosing a0=1/2, b0=1, t=sin²(u) gives the rank-four Abel relation. For all these
upper parameters in (0,1), iteration down to (1-z)^(-a) also proves positivity
for z<1. DLMF 16.5.1 supplies the separate Barnes formula for negative z.
The generator uses unregularized _5F4 with F(0)=1 and principal continuation
to z<1. Exact argument and coefficient identities are executable checks.

Checked the December 2023 DPVZ errata: the two corrections insert the central
character in arithmetic Hecke quadratics. They do not alter the analytic
definition or selected roster. No current-status claim about all inverse-pi
identities is made.

Live discovery on this run: cached authenticated catalogue including drafts,
screen.already_here, screen.already_asked and two expected numerical lookups.
No matching family found. T679 is a distinct rank-four normalized _4F3 family;
T680 is its mirror coordinate; T674 is rank-four weight-three local arithmetic.
T700/T701 mention Guillera but define elliptic dilogarithms, unrelated here.
Existing tags period, number theory, special functions and special values were
confirmed. The proposed new hypergeometric motive tag is absent and not added.
Claimed private draft T738; initial message cites #333 and contributes to #261.

The preceding screening report location supplied by #333,
numberdb/numberdb-runs:ideas/BATCH-2026-11-17T1231.md.gz, returned GitHub API 404
via the connected GitHub tool. The detailed #333 proposal and live #261 issue
were read instead; this access limitation does not establish anything about
the missing report's content. The shared dossier was initially empty.

All artifacts are in this directory. source-hashes.json records the downloaded
PDF, HTML, errata, live issues, catalogue and archived skill manifest hashes.
prepared-values.json and computation.jsonl preserve 150/200-digit checkpoints
and package versions. check_values.py is the builder's separate-formula check;
its evidence is not an independent critic's verdict. No sibling table is used
as numerical evidence. argument-selection.md defines every selected pair.

Pending family plan: the local L-polynomial proposal remains unbuilt by this
actor. A future companion should link T738 using the common (template,z)
identity, while retaining its rational-only domain and its separate quantity.
No Asai, bad-prime or global L-value computations were attempted here.

Skill: ../20261117T124903Z-build-skills/manifest.json (matching local/live bundle
889383d95f882353fc09d210af003abe6edcf0331f27c377d7acdabafa344208).