Follow the source behind the result.

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

"""v420 -- F_pole, the external side: the Koide source->pole transfer is a
ONE-LOOP ELECTROMAGNETIC effect, not an algebraic step.  This is the honest
"do the external physics" follow-up to v93 (which left the source->pole flow
time t~2.84 as an undetermined [P] generator) and to the clock No-Go (the
unitary/algebraic lens cannot reach a contraction rate).  It IDENTIFIES the
open piece as QED -- size, sign and scheme-dependence -- and keeps it firewalled
([C], external), NOT promoted.  Numerical (mpmath), Python-only (like v93/v371).

  [E] 1. KOIDE RESCALING INVARIANCE.  Q(m) = sum m / (sum sqrt m)^2 is homogeneous
         of degree 0: Q(c*m) = Q(m) for every c>0.  So any FLAVOR-UNIVERSAL
         running (a common factor on all three masses) leaves the Koide ratio
         EXACTLY fixed -- the source->pole shift is purely the FLAVOR-SPLIT
         (per-lepton) part of the running.
  [C] 2. THE GAP IS ONE-LOOP-EM-SIZED.  The tree (source) ladder gives
         2/3 - Q_src = 0.002203, which is the TFPT seed quantum phi0/24 = 0.002216
         to 0.6% AND the electromagnetic one-loop scale alpha/pi = 0.002323 to ~5%;
         the PDG pole sits AT 2/3 (gap 2.2e-6, the IR attractor).  So the
         source->pole gap is a one-loop EM size -- the transfer is QED.
  [C] 3. FLAVOR-SPLIT QED ESTIMATE (order + sign).  A standard 1-loop QED
         pole<->MSbar relation with the per-lepton log shifts the Koide by an
         EM-loop-order amount, and the pole (IR) is closer to 2/3 than the MSbar
         (UV) value -- i.e. the IR is the attractor (v82 direction).  The size is
         EM-loop; the exact value is scheme/scale-dependent.
  [O]/[C] 4. HONEST NEGATIVE / FIREWALL.  The EXACT flow time t~2.84 (v93)
         depends on the renormalisation scheme and the tree-scale identification
         -- external QED choices -- so it is NOT derived here, only sized.  The
         algebraic lens forces the RATE (2/3)^6 (v327/v82); this external EM
         physics supplies the source->pole flow.  F_pole stays [C], firewalled
         (v187): never re-pressed into the compiler.

VERDICT: the Koide source->pole transfer is identified as a one-loop EM effect
(size ~alpha/pi, IR-attractor direction), confirming F_transfer is genuinely
external -- the open theme is sharpened (QED), not closed.
"""
import mpmath as mp

from tfpt_constants import check, summary, reset, phi0

mp.mp.dps = 40

ALPHA = 1 / mp.mpf('137.0359992168')          # the TFPT-derived alpha (v3)
TWO3 = mp.mpf(2) / 3
# PDG 2024 pole masses (MeV)
ME, MMU, MTAU = mp.mpf('0.51099895069'), mp.mpf('105.6583755'), mp.mpf('1776.93')
MZ = mp.mpf('91187.6')


def koide(m1, m2, m3):
    return (m1 + m2 + m3) / (mp.sqrt(m1) + mp.sqrt(m2) + mp.sqrt(m3))**2


def run():
    reset()
    print("v420 F_pole external: the Koide source->pole transfer is one-loop EM")

    me, mmu, mtau = mp.mpf(16) / 7 * phi0**5, mp.mpf(4) / 3 * phi0**3, mp.mpf(7) / 6 * phi0**2
    q_src = koide(me, mmu, mtau)
    q_pole = koide(ME, MMU, MTAU)
    gap = TWO3 - q_src

    # ---- 1. Koide rescaling invariance (exact) ----
    inv = all(abs(koide(c * me, c * mmu, c * mtau) - q_src) < mp.mpf('1e-30')
              for c in (mp.mpf('0.137'), mp.mpf('3.7'), mp.mpf('1e6')))
    check("KOIDE RESCALING INVARIANCE [E]: Q(c*m)=Q(m) for all c>0 (degree-0 "
          "homogeneous) -- flavor-UNIVERSAL running leaves Koide exactly fixed, "
          "so the source->pole shift is purely the flavor-SPLIT part",
          inv)

    # ---- 2. the gap is one-loop-EM-sized ----
    check("ONE-LOOP-EM SIZE [C]: 2/3 - Q_src(tree) = %.6f = phi0/24 (%.6f) to "
          "0.6%% and = alpha/pi (%.6f, one-loop EM) to ~5%%; the PDG pole sits "
          "AT 2/3 (gap %.1e, the IR attractor) -- the source->pole gap is a "
          "one-loop EM size"
          % (float(gap), float(phi0 / 24), float(ALPHA / mp.pi), float(TWO3 - q_pole)),
          abs(gap / (phi0 / 24) - 1) < mp.mpf('0.01')
          and mp.mpf('0.85') < gap / (ALPHA / mp.pi) < mp.mpf('1.0')
          and abs(TWO3 - q_pole) < mp.mpf('1e-5'))

    # ---- 3. flavor-split 1-loop QED estimate (order + sign) ----
    def mbar(mpole):     # standard 1-loop QED pole->MSbar (per-lepton log)
        return mpole * (1 - (ALPHA / mp.pi) * (1 + mp.mpf(3) / 4 * mp.log(MZ**2 / mpole**2)))
    q_msbar = koide(mbar(ME), mbar(MMU), mbar(MTAU))
    shift = q_msbar - q_pole
    check("FLAVOR-SPLIT QED ESTIMATE [C]: a 1-loop QED pole<->MSbar relation "
          "(per-lepton log) shifts Koide by |dQ|/(alpha/pi) = %.2f (EM-loop "
          "order), and the pole (IR) is closer to 2/3 than the MSbar (UV) -- the "
          "IR is the attractor (v82 direction); exact value scheme-dependent"
          % float(abs(shift) / (ALPHA / mp.pi)),
          mp.mpf('0.1') < abs(shift) / (ALPHA / mp.pi) < mp.mpf('3')
          and abs(TWO3 - q_pole) < abs(TWO3 - q_msbar))

    # ---- 4. honest negative / firewall ----
    def rho(q):
        return (3 * q - 2) / (5 - 3 * q)
    mu = (mp.mpf(2) / 3)**6
    t_flow = mp.log(rho(q_pole) / rho(q_src)) / mp.log(mu)
    check("HONEST NEGATIVE / FIREWALL [O]/[C]: the EXACT flow time t = %.3f "
          "(v93) is scheme/scale-dependent external QED -- NOT derived, only "
          "sized; the algebraic lens forces the RATE (2/3)^6 (v327/v82), this "
          "external EM physics supplies the flow; F_pole stays [C], firewalled "
          "(v187)" % float(t_flow),
          mp.mpf('2.7') < t_flow < mp.mpf('3.0')
          and abs(t_flow - mp.nint(t_flow)) > mp.mpf('0.05'))

    return summary("v420 F_pole external (source->pole = one-loop EM; rate "
                   "forced, flow external [C])")


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

c9cc784c6a27030600d512463df8bb6d76ae1c8b8af16d2d9e94cfb605c730a6

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.