Read here
Web reading route
A concise, phone-friendly path through the object, the eight parts and the exact research boundary.
Start the route ↓Volume I · Web reading route
Triangular-n Postmaster.
From the triangular primitive to the completed source frontier.
Begin with a familiar triangle, follow its reflection coordinate into the completed zeta source, and stop exactly where the proved part stops. This page is the approachable layer; the Reading Volume and Technical Dossier remain one click deeper.
DOCUMENT v6.3
OCTOBER 2026
RH STATUS: OPEN
Choose the depth
One work, three reading layers.
The HTML route gives the shape of the argument. The Reading Volume carries the continuous mathematics. The Dossier carries the full proofs, certificates, countermodels and empirical context.
Read continuously
Reading Volume
108 pages, sections R1-R25. The argument, equations, figures, status key and final summary in one document.
Open the PDF ↗Peel in deep
Technical Dossier
301 pages of proofs, finite certificates, comparison models, failed shortcuts, source calculations and empirical context.
Open the Dossier ↗The v6.3 source package (masters, figures, build and check scripts; 27.9 MB) is on the Zenodo record. Every earlier edition is in the Zenodo version series.
Calibration · since v5.2
A twin that passes 16,588 rungs.
Since version 5.2 the volume carries a calibration. The Davenport–Heilbronn function shares the reflection symmetry of completed zeta, F(s) = F(1 − s), but has a different Gamma factor and no Euler product. On a finite list of its zeros, sampled scans find every Widder rung positive through W16588 and detect a negative at W16589. At the three tested centres its Löwner sections stay positive through a hundred nodes; they first fail at exactly 105, 106 and 107 nodes. So a long finite rung prefix does not by itself rule out off-line zeros. No status changed. RH remains open.
The twin evidence first shipped with the v5.2 source package (Zenodo 10.5281/zenodo.23004335). See the result drawn in 51 Twins, and the numbers in the twin calibration report.
Start here
The triangle is the entrance, not the conclusion.
Triangular numbers count familiar things: stacked objects, handshakes and the sum 1 + 2 + … + n. After centering the same primitive at its half-step, its reflection symmetry continues to the complex coordinate used by the volume.
Tn = n(n + 1) / 2
→
q = s(1 − s) = −2Ts−1
The fold sends s and 1 − s to the same point. It does not silently merge conjugate points, and it does not assume RH. That distinction is one of the clarifications made explicit in v5.1.
The next step is a signed Pascal row. Each row gives an exact finite formula for one Widder rung. The full question is whether every rung is nonnegative at every positive scale. The phrase every rung is where the open problem lives.
Root to frontier
Read the argument in eight parts.
Each part stays readable at this level, then offers the exact Reading Volume page and a relevant Dossier entrance.
IReading pp. 10-25
The triangular coordinate
Build the discrete primitive, its reflection, reciprocal moments, the Euler-constant bridge and the half-step coordinate. This is the arithmetic root of the later analytic construction.
Carry forward: one centered coordinate can hold several exact identities without making those identities the same theorem.
IIReading pp. 26-30
From the half-step to the complex fold
Continue the centered triangular symmetry into q = s(1 − s). Reflection partners share a folded point; conjugation remains visible as a separate operation away from the critical line.
Carry forward: the fold removes the branch label, not the information needed to recognize critical-line membership.
IIIReading pp. 31-42
Completion and the folded resolvent
The completed function is written in the folded coordinate, then organized as a resolvent, a heat trace and related positive representations. The volume states exactly how reflection pairs and multiplicity are counted.
Carry forward: positivity for a positive real folded node does not automatically extend to an individual nonreal kernel.
IVReading pp. 43-63
The Widder ladder and sector geometry
Every rung becomes one finite signed-Pascal row. The volume proves the first rung analytically, retains finite source certificates for the next five, and separates that route from the much longer verified-height prefix.
Carry forward: a finite prefix can be rigorous and enormous while still not supplying an all-order theorem.
VReading pp. 64-81
Li data, matrices and source energy
The same folded data is read through Li coefficients, Löwner matrices, Gamma and prime blocks, Bernstein rows, Laguerre energy, cutoffs and exact-divisor residuals.
Carry forward: exact translations between forms do not make an unproved sign estimate disappear.
VIReading pp. 82-90
Translation defects and reflection contours
Here the volume isolates what an off-line point would leave behind: contour, heat and Green-function defects. These are exact detectors, but the source-side signs needed to close the argument remain separate.
Carry forward: detecting the shape of a failure is not the same as proving that no failure exists.
VIIReading pp. 91-96
Results and the research boundary
Results are restated with their quantifiers, then each surviving route is stopped at its first unsupported line. No unproved reduction between the matrix, energy, divisor and rung routes is assumed.
Carry forward: the next independent source rung is W7; the full criterion still requires every rung at every positive scale.
VIIIReading p. 97
What this approach cannot do
The no-go ledger records shortcuts already ruled out: finite-prefix stopping rules, coarse large-sieve closure, selected Gamma sublattices and other tempting replacements for the actual signed arithmetic remainder.
Carry forward: a failed shortcut narrows the corridor; it is neither evidence for RH nor evidence against it.
The final page of the argument
Object, results, remaining line.
The volume ends by placing the coordinate, the proven finite results and the missing uniform source positivity in one summary.
What the fold preserves
Reflection is folded. Conjugation is not.
One reflection orbit {ρ, 1 − ρ} gives one folded node qρ. If a zero were off the critical line, its conjugate reflection orbit would give a second, conjugate folded node.
This is why the volume is careful about the index set and multiplicity before it sums anything. Reciprocal and sign changes of a nonzero folded coordinate are reversible; taking only a real part is a further observation and can lose information.
Where the work stands
Established results on one side. Full quantifiers on the other.
The boundary is mathematical, not rhetorical. Every result keeps the range and assumptions that make it true.
Established in the volume
The construction and its finite reach
- The triangular primitive, reflection fold, resolvent identities and signed-Pascal rung formulas are exact.
- The first source rung is proved analytically.
- Source rungs W2-W6 retain their rigorous finite certificates from the v5.0 build.
- A separate verified-height theorem gives a much longer, but still finite, positive rung prefix: every rung through W4,772,432,051,943 with the published zero-free region as a second input (13B.3).
- On the same channel, with the published zero count as a second input, Gn,k(x) > 0 for every k ≥ 0, every 0 < x ≤ H2/106 and every n ≤ 4,500,262,999,199 (13C.4).
- The no-go results rigorously exclude their stated shortcuts.
Still open
The uniform source statement
- The next independent source proof, beginning at W7, is not complete.
- No source-side proof establishes every rung for every positive scale.
- The all-node Gamma-minus-prime matrix inequality remains unproved.
- The corresponding all-degree energy and exact-divisor bounds remain open on their stated routes.
- Δ < 1, equivalently a zero-free half-plane ℜs > θ0 with θ0 < 1, is not proved. An external claim with θ0 = 7/8 is read, not verified, and used only as a stated assumption (R13C).
- The Riemann hypothesis is not proved.
Exact criterion
RH ⇔ Wk(x) ≥ 0 for every k ≥ 1 and every x > 0
Current edition
What v6.3 carries.
Version 6.3 brings the volume up to date with the companion paper Reading the Fold (v1.3) and the notation the two now share. No v6.2 result, number, status or equation tag changes, and the Riemann hypothesis remains open.
The half index and the triangular product
At the vertex, T(−½) = −1/8 (TDP.5). The product ∏(1 + iy) has coefficients ek(n), and at n = −½ they are 1, −1/8, 1/128, 5/1024, … (TDP.6). Its interior coefficients have gcd n + 1 when n + 1 is prime and 1 otherwise: the successor gate again (TPG.1–TPG.3).
The reciprocal-triangular law
Halving Σ 1/Tn = 2 gives a probability law. Its entropy is −2Σζ′(2k) = log 2 + 2Σ(ζ(2k) − 1)/(2k − 1) = 2.0462774528… nats (4.6), the entropy of the Lüroth map. On the fold, q = s(1 − s) is a coin’s variance.
π on the odd squares, and 990 in scope
Wallis’s product reads π = 4∏ 8Tn/(8Tn + 1) (PR.4b). Dossier §170 now says what 990 = T44 does not mean: it is one of six triangular products of three consecutive integers below 2·106, and in the zeta ladder 990 comes from Euler’s formula.
One letter, one meaning
T(z) at every non-integer argument, x = s(s − 1) = −q, N0(t) for the zero count and τn for the Taylor coefficients of (12.9), as in Reading the Fold. Two sentences of the proof prose of Lemma 96A.1 are corrected.
Publication details
Volume I, v6.3 · October 2026
This site hosts v6.3, published 20 October 2026. Its Zenodo record is 10.5281/zenodo.23510962; v6.2 is 10.5281/zenodo.23251531. The concept DOI below is the durable link for the version series.
Huckstead, Jeffery. Triangular-n Postmaster: From the Triangular Primitive to the Completed Source Frontier. Volume I - The Reading Volume, v6.3. Cerebral Graphix, October 2026. DOI: 10.5281/zenodo.23510962. Version series: 10.5281/zenodo.21968915. ORCID: 0009-0007-0234-2177.