Follow the source behind the result.

The registered Python source is shown as text. Displaying it does not execute the code.

#!/usr/bin/env python3
"""v624 -- EXTREV.LATTICE.01: the external-review lattice audit -- three
claimed derivations machine-checked: the FOUR-MARK CLOSURE (d = 4 forced three
ways under the D_{d+1} + A_{d-1} glue architecture), the ANCHOR BYTECODE
(one cubic polynomial and a two-state recursion carry the discrete
source code), and the LORENTZ CONGRUENCE (the prime-front determinant
form and the cover polarization lattice are the SAME rational quadratic
form, index 6).  Honest negative included: the review's follow-up B-test
is ill-posed as stated.

An external review (2026-08-02) proposed three new derivations plus an
RH proof architecture.  Per the audit workflow, every checkable claim is
machine-verified before adoption:

  (A) THE FOUR-MARK CLOSURE [E, conditional]: generalize the glue to
      D_{d+1} + A_{d-1}.  Three INDEPENDENT selectors each force d = 4:
      (A1) disc-group matching: disc(A_{d-1}) = Z_d is cyclic; a
      cyclic-x-cyclic glue needs disc(D_{d+1}) cyclic, i.e. d even, and
      Z_d = Z_4 forces d = 4 (census d = 3, 4, 5, 8 with negative
      controls);
      (A2) the even-glue norm equation (d+1)/4 + (d-1)/d = 2 has the
      UNIQUE positive solution d = 4, and no other even value 2k
      (k <= 10) admits an integer solution;
      (A3) the Pythagorean selector (d-1)^2 + d^2 = (d+1)^2 holds for
      positive d exactly at d = 4 -- the (3, 4, 5) triple of (cycles,
      marks, carrier) is the unique Pythagorean closure.
      Consequences: (1,1,2) is the UNIQUE 3-part partition of 4, and
      c_3 = 1/(2 e_1 pi) = 1/(8 pi): under the named architecture the
      two-input presentation reduces to ONE boundary degree d = 4
      (plus pi).  Honest typing: conditional on the glue architecture
      (genus-0 boundary, A-side cycles, D-side functions, cyclic even
      unimodular glue); the d = 4 special case (5/4 + 3/4 = 2) was
      already exact in the corpus (v47-era selection).

  (B) THE ANCHOR BYTECODE [E, consolidation]: chi_a(t) = (t-1)^2(t-2)
      = t^3 - 4t^2 + 5t - 2 carries (e_1, e_2, e_3) = (4, 5, 2) =
      (marks, carrier, sheet involution); the power sums p_n = 2 + 2^n
      satisfy the two-state recursion p_{n+2} = 3 p_{n+1} - 2 p_n
      (p_0 = 3, p_1 = 4); Newton identities recover the e_k; and
      p_1 p_2 p_3 = 240, p_4 - p_3 = 8, 240 + 8 = 248 = dim E_8.
      The operator realization H = 1 + V(V-1)/2 with char chi_a is
      already machine-checked (v602) -- this packages the discrete
      source as: d = 4 -> chi(t) -> recursion -> E_8.

  (C) THE LORENTZ CONGRUENCE [E + honest negative]:
      (C1) det X = (1/2) y^T J_det y with y = (X11, X22, X12) and
      J_det = [[0,1,0],[1,0,0],[0,0,-2]] (the prime-front determinant
      form; det 2, signature (1,2));
      (C2) THE CONGRUENCE: P = [[3,0,0],[3,0,2],[-1,1,-1]] with
      det P = -6 satisfies P^T J_det P = J_fix EXACTLY (J_fix the v604
      integer polarization form, det 72): the two Lorentz lattices are
      the SAME rational quadratic form, the cover an index-6 sublattice;
      72/2 = 36 = 6^2 and 6 = p_2(a) (the A_3 positive-root count);
      (C3) [X, must-fail honest] the review's follow-up test
      "integer R with R^T J_det R = B, |det R| = 2" is ILL-POSED as
      stated: B = M^dagger J (v602) is COMPLEX symmetric (det 8
      verified) -- no real congruence can produce it;
      (C4) [E, partial positive] on the reformulated field level the
      third symmetric-Gauss diagonal ratio IS an exact Q(omega) square
      ((1 - sqrt3 i)^2 = -2 - 2 sqrt3 i); the full Q(omega)-congruence
      (Hasse invariants) and the CANONICITY of P (derivation from the
      cover operators) are the named open steps.

Verdict enums (frozen): LATTICE-AUDIT-PASSED (all pass),
AUDIT-FAILS, MIXED.

FIREWALL: no marker changes; AX.P2.01 keeps its typing (the closure is
conditional on the glue architecture); no RH statement.

PROVENANCE: discovery probe external_lattice_audit_probe.py
(2026-08-02, 14/14, verdict LATTICE-AUDIT-PASSED); external review
2026-08-02 (three derivations + RH architecture; Suzuki arXiv
2606.09096 and the numerical realization arXiv 2607.24830 verified by
web lookup).

Python-only (sympy, exact), counted per GATE.WOLFRAM.02.
"""

from sympy.utilities.iterables import partitions

import sympy as sp

CHECKS = []


def check(name, ok, detail=""):
    CHECKS.append((name, bool(ok)))
    print("[%s] %s%s" % ("PASS" if ok else "FAIL", name,
                         (" -- " + detail) if detail else ""))


# ================================================================== A
print("=" * 72)
print("A: the four-mark closure (d = 4 forced three ways)")
print("=" * 72)

d = sp.symbols("d", positive=True)

census_ok = True
for dd in (3, 4, 5, 8):
    cyclicD = (dd + 1) % 2 == 1          # disc(D_{d+1}) cyclic iff d even
    group_match = cyclicD and dd == 4    # Z_d = Z_4
    norm = sp.Rational(dd + 1, 4) + sp.Rational(dd - 1, dd)
    norm_even_int = norm.is_integer and norm % 2 == 0
    if (dd == 4) != (group_match and norm_even_int):
        census_ok = False
check("A1 disc-group census (d = 3, 4, 5, 8): disc(A_{d-1}) = Z_d; "
      "disc(D_{d+1}) cyclic iff d even; the cyclic glue match Z_d = Z_4 "
      "AND the even integer glue norm hold EXACTLY at d = 4 (negative "
      "controls at 3, 5, 8)", census_ok)

sols = sp.solve(sp.Eq((d + 1) / sp.Integer(4) + (d - 1) / d, 2), d)
k_census = True
for k in range(2, 11):
    ss = sp.solve(sp.Eq((d + 1) / sp.Integer(4) + (d - 1) / d, 2 * k), d)
    if any(s.is_integer and s > 1 for s in ss):
        k_census = False
check("A2 the even-glue norm equation (d+1)/4 + (d-1)/d = 2 has the "
      "UNIQUE positive solution d = 4, and no other even target 2k "
      "(k = 2..10) admits an integer d > 1", sols == [4] and k_census)

pyth = sp.solve(sp.Eq((d - 1) ** 2 + d ** 2, (d + 1) ** 2), d)
check("A3 the Pythagorean selector (d-1)^2 + d^2 = (d+1)^2 holds exactly "
      "at d = 4: the (3, 4, 5) = (cycles, marks, carrier) triple is the "
      "unique Pythagorean closure", pyth == [4])

P3 = [tuple(sorted(sum(([kk] * v for kk, v in pp.items()), [])))
      for pp in partitions(4) if sum(pp.values()) == 3]
check("A4 (1, 1, 2) is the UNIQUE 3-part partition of 4 (the anchor is "
      "forced given e_1 = 4), and c_3 = 1/(2 e_1 pi) = 1/(8 pi)",
      P3 == [(1, 1, 2)]
      and sp.Rational(1, 2 * 4) / sp.pi == sp.Rational(1, 8) / sp.pi)

check("A5 [C] honest typing: the closure is CONDITIONAL on the glue "
      "architecture (genus-0 boundary with d marks, A_{d-1} cycle side, "
      "D_{d+1} function side, cyclic even unimodular glue); under it the "
      "two-input presentation (c_3, g_car) reduces to ONE boundary degree "
      "d = 4 plus pi; the residue moves to: why the physical boundary "
      "realizes the minimal cyclically-glueable genus-0 divisor", True)

# ================================================================== B
print("=" * 72)
print("B: the anchor bytecode")
print("=" * 72)

t = sp.symbols("t")
chi = sp.expand((t - 1) ** 2 * (t - 2))
check("B1 chi_a(t) = (t-1)^2(t-2) = t^3 - 4t^2 + 5t - 2 carries "
      "(e_1, e_2, e_3) = (4, 5, 2) = (marks, carrier, sheet involution)",
      chi == t ** 3 - 4 * t ** 2 + 5 * t - 2)


def p(n):
    return 2 + 2 ** n


check("B2 the power sums p_n = 2 + 2^n satisfy the two-state recursion "
      "p_{n+2} = 3 p_{n+1} - 2 p_n with (p_0, p_1) = (3, 4)",
      all(p(n + 2) == 3 * p(n + 1) - 2 * p(n) for n in range(8))
      and p(0) == 3 and p(1) == 4)

e1 = p(1)
e2 = (p(1) ** 2 - p(2)) // 2
e3 = (p(3) - e1 * p(2) + e2 * p(1)) // 3
check("B3 Newton identities recover the anchor coefficients "
      "(e_1, e_2, e_3) = (4, 5, 2) from the recursion alone",
      (e1, e2, e3) == (4, 5, 2))

check("B4 p_1 p_2 p_3 = 4*6*10 = 240 and p_4 - p_3 = 8, with "
      "240 + 8 = 248 = dim E_8 (the operator realization H = 1 + V(V-1)/2 "
      "with char chi_a is v602-checked)",
      p(1) * p(2) * p(3) == 240 and p(4) - p(3) == 8
      and p(1) * p(2) * p(3) + p(4) - p(3) == 248)

# ================================================================== C
print("=" * 72)
print("C: the Lorentz congruence (and the honest negative)")
print("=" * 72)

Jdet = sp.Matrix([[0, 1, 0], [1, 0, 0], [0, 0, -2]])
Jfix = sp.Matrix([[16, 2, 4], [2, -2, 2], [4, 2, -2]])
P = sp.Matrix([[3, 0, 0], [3, 0, 2], [-1, 1, -1]])

X11, X22, X12 = sp.symbols("X11 X22 X12")
y = sp.Matrix([X11, X22, X12])
def inertia(F):
    evs = [complex(sp.N(ev, 30)).real for ev in F.eigenvals(multiple=True)]
    return (sum(1 for e in evs if e > 1e-12), sum(1 for e in evs if e < -1e-12))


check("C1 det X = (1/2) y^T J_det y exactly (y = (X11, X22, X12)); "
      "det J_det = 2; both forms have Lorentz signature (1,2)",
      sp.simplify(sp.Rational(1, 2) * (y.T * Jdet * y)[0, 0]
                  - (X11 * X22 - X12 ** 2)) == 0
      and Jdet.det() == 2
      and inertia(Jdet) == (1, 2) and inertia(Jfix) == (1, 2))

check("C2 THE CONGRUENCE: P^T J_det P = J_fix EXACTLY with det P = -6: "
      "the prime-front determinant form and the v604 cover polarization "
      "lattice are the SAME rational quadratic form, the cover an "
      "index-6 sublattice; 72/2 = 36 = 6^2 and 6 = p_2(a)",
      sp.simplify(P.T * Jdet * P - Jfix) == sp.zeros(3, 3)
      and P.det() == -6 and Jfix.det() == 72
      and sp.Rational(72, 2) == 36 and p(2) == 6)

# C3: B = M^dagger J (v602 verbatim) is complex symmetric -- the literal
# integer-R congruence test is ill-posed
M0 = sp.Matrix([[-1, (1 + sp.sqrt(3) * sp.I) / 2, (1 - sp.sqrt(3) * sp.I) / 2],
                [0, 0, 1],
                [0, 1, 0]])
J = sp.Matrix([[0, 1 + sp.sqrt(3) * sp.I, -1 + sp.sqrt(3) * sp.I],
               [1 - sp.sqrt(3) * sp.I, 2, 1 + sp.sqrt(3) * sp.I],
               [-1 - sp.sqrt(3) * sp.I, 1 - sp.sqrt(3) * sp.I, 0]])
B = sp.expand(sp.simplify(M0.conjugate().T * J))
B_complex = any(sp.simplify(sp.im(sp.expand_complex(e))) != 0 for e in B)
check("C3 [X, must-fail honest] the review's follow-up test (integer R "
      "with R^T J_det R = B, |det R| = 2) is ILL-POSED as stated: "
      "B = M^dagger J is COMPLEX symmetric (det B = 8 verified) -- no "
      "real congruence can produce a complex form",
      sp.simplify(B - B.T) == sp.zeros(3, 3)
      and sp.nsimplify(sp.simplify(B.det())) == 8 and B_complex)

# C4: partial positive on the reformulated field level
ratio3 = sp.simplify(sp.expand_complex((1 - sp.sqrt(3) * sp.I) ** 2))
check("C4 [E, partial] on the reformulated Q(omega) level the third "
      "symmetric-Gauss diagonal ratio is an EXACT square: "
      "(1 - sqrt3 i)^2 = -2 - 2 sqrt3 i; the full Q(omega)-congruence "
      "(Hasse invariants) and the CANONICITY of P (derivation from the "
      "cover operators) are the named open steps",
      sp.simplify(ratio3 - (-2 - 2 * sp.sqrt(3) * sp.I)) == 0)

check("C5 [C] the reading: the congruence is a REAL new bridge (prime "
      "front <-> Hodge geometry) at the form level; its significance "
      "upgrade requires P canonical (from real structure, polarization "
      "or period basis) and the window vectors landing in the positive "
      "Hodge chamber -- preregistered as PRIME.COVERLORENTZ follow-ups; "
      "no positivity claim, no RH statement", True)

# ================================================================== summary
print("=" * 72)
n_pass = sum(1 for _, ok in CHECKS if ok)
print("%d/%d checks passed" % (n_pass, len(CHECKS)))
if n_pass == len(CHECKS):
    print("ALL CHECKS PASSED")
    print("VERDICT: LATTICE-AUDIT-PASSED -- the four-mark closure forces d = 4")
    print("three independent ways (conditional on the glue architecture), the")
    print("anchor bytecode packages the discrete source exactly, and the")
    print("prime-front/cover Lorentz congruence is exact (index 6); the")
    print("review's B-test is retyped as ill-posed with the honest")
    print("reformulation named.")
else:
    print("SOME CHECKS FAILED")


def run():
    """run_all.py entry point; the checks execute at import time above."""
    return len([1 for _, ok in CHECKS if not ok])


if __name__ == "__main__":
    raise SystemExit(1 if run() else 0)

71d0e6fbabad7f3d2187d708638371a2df5722f5a9d5eb80d851e205e577cc9b

The public registry includes 116 document views, 6,693 section records and 985 physical LaTeX sources. Active publication documents, research material and archives retain their recorded roles. This is source navigation, not a new semantic review or proof run.