The discreteness map is complete, and the frame-independent flank closes by weak PREDICTION, not weak data: FIRAS and cold cosmic hydrogen exclude unsuppressed Planckian momentum diffusion ('swerves') by 65–111 orders of magnitude, but causal-set theory derives no value for the diffusion constant κ — candidate scalings span >100 orders, the massive and photon sectors cross their bounds at different suppression powers (n* = 3.8 vs 2.0), and the next rung down is unreachable by 10³–10³⁰. No feasible swerve measurement distinguishes a Lorentz-invariant-discrete universe from a continuum.

S (the verdict computation)
==============================================================================
There is NO derived kappa in causal set theory: DHS introduce it as a free
phenomenological parameter.  'Natural' can only mean dimensional analysis, and
dimensional analysis does not close on a unique answer -- that non-uniqueness is
itself a finding.  Candidate family (massive sector, m = m_H, r = m/m_Planck):

    kappa_n = (m c)^2 / t_Planck * r^n ,  n = degree of Planck suppression
    plus the particle-blind minimal form kappa_P = (m_Planck c)^2 / t_Planck.
    r = m_H/m_Planck = 7.6894e-20;   (m_H c)^2/t_Planck = 4.6691e+06 kg^2 m^2 s^-3
    candidate                                           value  bound/candidate  status
    kappa_P  (pure Planck, no particle scale)       7.897e+44       6.719e-112  EXCLUDED
    kappa_0  (unsuppressed, O(1) kick/t_Pl)         4.669e+06        1.136e-73  EXCLUDED
    kappa_2  (quadratic suppression r^2)            2.761e-32        1.922e-35  EXCLUDED
    kappa_3  (cubic suppression r^3)                2.123e-51        2.499e-16  EXCLUDED
    kappa_4  (quartic suppression r^4)              1.632e-70        3.250e+03  allowed (unreachable)
    crossing power: the hydrogen bound slices the family at n* = 3.82
    conservative-variant audit: status flips = 0 of 5 candidates;
    n* moves 3.82 -> 3.66.  The verdict does not depend on the gas choice.

Photon sector (E = k_B T_CMB, r_g = E/E_Planck):
    r_g = 1.9237e-32;   (k_B T)^2/t_Planck = 2.6265e-02 J^2 s^-1
    candidate                                     value  bound/candidate  status
    kappa_g0 (unsuppressed)                   2.626e-02        2.728e-65  EXCLUDED
    kappa_g1 (linear suppression r_g)         5.053e-34        1.418e-33  EXCLUDED
    kappa_g2 (quadratic r_g^2)                9.720e-66        7.372e-02  EXCLUDED
    kappa_g3 (cubic r_g^3)                    1.870e-97        3.832e+30  allowed (unreachable)
    crossing power: the FIRAS bound slices the photon family at n* = 2.04
------------------------------------------------------------------------------
READING:
  - Unsuppressed Planckian swerves are excluded by ~9e+72 (massive)
    and ~4e+64 (photon).  The particle-blind form by ~1e+111.
  - But the candidate family spans 76 orders of magnitude as n runs 0->4
    (115 orders including the pure-Planck form), and the two sectors cross
    their bounds at DIFFERENT powers (3.8 vs 2.0):
    there is no unique 'natural' kappa for the data to aim at.
  - One more power of suppression (n=4 massive, n=3 photon) drops the
    prediction to ~2e-70 / ~2e-97 -- below the bounds by
    3.3e+03x / 3.8e+30x.
  - To even reach the massive n=4 rung, a lab would need heating-rate
    sensitivity ~8.2e-23 K/s (Rb): 12.1 orders past 100 pK/s;
    no experiment on any roadmap reaches a rung the theory actually owns.
==============================================================================
5. REGISTRY CROSS-REFERENCES (current statements, re-read from disk)
==============================================================================
  USR-2026-0041 [provisional]: E_QG > 1.4e19 GeV, of order Planck — scales as 1/Δt on an unquantified 1 s window
  USR-2026-0037 [provisional]: no discriminator — recorded as framing, not evidence
  USR-2026-0038 [provisional]: a recomputed bound (3.4e123 bits) — nothing was measured against it
  USR-2026-0070 [provisional]: NO discriminator in the wing we built; the one live prediction is systematics-limited
==============================================================================
6. VERDICT
==============================================================================
NO DISCRIMINATOR IN PRACTICE for Lorentz-invariant discreteness.

The flank is not closed by weakness of data -- the data are spectacular: FIRAS
alone crushes unsuppressed Planckian photon diffusion by ~65 orders of
magnitude, and cosmic hydrogen crushes the massive sector by ~73.  The flank is
closed by weakness of PREDICTION: causal set theory derives no value of kappa.
The dimensional-analysis candidates span >100 orders, the two particle sectors
cross their bounds at different suppression powers, and every rung at or below
one extra power of m/m_Planck is unreachable by any conceivable experiment.
A null at ANY achievable precision is accommodated by a smaller kappa; a
detection at any precision has no rung to confirm.  Same epistemic bin as
USR-2026-0037/0038 (and the emergent-gravity wing, USR-2026-0070): an
observable exists, but no measurement outcome can move the hypothesis.

THE DISCRETENESS MAP IS NOW COMPLETE for this project:
  - FRAME-DEPENDENT discreteness (preferred frame, granular dispersion, lattice
    anisotropy, clock/grid/float artifacts): EXCLUDED BY DATA -- our five null
    seams plus the reproduced GRB dispersion bound (USR-2026-0041, current
    corrected statement above: order-Planck reach, 1/Delta-t caveat).
  - FRAME-INDEPENDENT discreteness (causal sets): predicts no dispersion, so it
    evades all of the above; its swerve diffusion is bounded (this test) tens of
    orders below every unsuppressed Planckian scaling, yet remains UNFALSIFIABLE
    IN PRACTICE because kappa is a free parameter with no derived magnitude.
  - The causal-set flank has exactly TWO known observational handles:
    (1) swerves -- this test: no discriminator in practice;
    (2) everpresent Lambda (Ahmed-Dodelson-Greene-Sorkin) -- tested separately
        in this wave.  No third handle is established in the literature.

Discriminator test applied to this result: a computed universe implementing
Lorentz-invariant discreteness and an uncomputed continuum universe give the
SAME answer to every feasible swerve measurement (namely: nothing, at current
and any plausible future sensitivity).  That is the definition of no
discriminator.  Recorded as a first-class null-of-method, not a defeat.

HEADLINES:
  H1  kappa(cosmic H, adopted)  < 5.306e-67 kg^2 m^2 s^-3  (conservative variant 5.306e-64)
  H2  kappa(ultracold Rb-87)    < 1.992e-58 kg^2 m^2 s^-3
  H3  kappa_gamma(FIRAS mu)     < 7.165e-67 J^2 s^-1  = 7.972e-84 kg^2 m^2 s^-3
  H4  strongest: FIRAS on the photon constant; cosmic H on the massive constant
  H5  Planckian candidates: massive 7.90e+44 (pure), 4.67e+06 (n=0) ... 1.63e-70 (n=4);
      photon 2.63e-02 (n=0) ... 1.87e-97 (n=3)  [kg^2 m^2 s^-3 / J^2 s^-1]
  H6  exclusion ratios: massive n=0 by 8.8e+72, n=2 by 5.2e+34, n=3 by 4.0e+15; n=4 allowed by 3.3e+03
      photon n=0 by 3.7e+64, n=1 by 7.1e+32, n=2 by 1.4e+01; n=3 allowed by 3.8e+30
  H7  crossing powers: n* = 3.82 (massive) vs 2.04 (photon) -- no consistent natural rung
  H8  VERDICT: NO DISCRIMINATOR IN PRACTICE (bin of USR-2026-0037/0038)
  H9  DISCRETENESS MAP COMPLETE: frame-dependent = excluded by data;
      frame-independent = unfalsifiable in practice (swerves) + everpresent Lambda (separate test)
==============================================================================
END TEST 24

This script needs data files, a heavy dependency,
or more time than a browser tab should take.
Download it from /data and run it locally.