Example

Find where Haraux's two collision times start

Compute the events of the paper's scenario exactly in both frames and see why two different times are not a contradiction.

A 2024 paper by A. Haraux derives two collision times for one scenario, D/(c + v) and D/c, and argues that their difference is a contradiction in the mathematics of special relativity. Two numbers that differ contradict each other only if they answer the same question. This program computes the events exactly, at v = 3/5 c, and asks where each time starts.

Run it

Save this as haraux_events.py and run it with Python 3.11 or later. It uses the standard library only.

"""Haraux's two collision times, exactly (standard library only).

Units c = 1, D = 1. The photon is emitted at x = 0 at S-time 0 (event E), towards
M, which is at x = D at S-time 0 (event A) and moves at -v. They meet at event B.
M's rest frame S' is reached by the Lorentz transformation for velocity -v.
"""
from fractions import Fraction as F

v = F(3, 5)
gamma = F(5, 4)                      # 1 / sqrt(1 - v^2), exact for v = 3/5
assert gamma * gamma * (1 - v * v) == 1


def to_m_frame(t, x):                # S -> S'
    return gamma * (t + v * x), gamma * (x + v * t)


E = (F(0), F(0))
A = (F(0), F(1))
B = (1 / (1 + v), 1 / (1 + v))       # the photon reaches M at S-time D / (c + v)
for name, ev in (("E: photon emitted", E), ("A: M at S-time 0", A), ("B: photon meets M", B)):
    print(f"{name:20} S {tuple(map(str, ev))!s:14} S' {tuple(map(str, to_m_frame(*ev)))}")

print()
print(f"E and A simultaneous in M's frame? {to_m_frame(*E)[0] == to_m_frame(*A)[0]}")
x_m = gamma * 1                      # M rests at x' = gamma * D in S'
print(f"photon speed in M's frame: {x_m / to_m_frame(*B)[0]}")

t1 = B[0] - E[0]                     # T1 = D/(c + v): from E to B
s_star = (gamma - 1 / gamma)         # the S'-slice on which the photon is alpha*D from M
t_start = gamma * (s_star - v * x_m) # S-time of M's event on that slice
t2 = B[0] - t_start                  # T2: from that event to B
print()
print(f"T1 = D/(c+v) = {t1}, starting at E (S-time 0)")
print(f"T2 = {t2} = D/c? {t2 == 1}, starting at M's event at S-time {t_start}")
print(f"the two start events differ by {t2 - t1} in S-time; T2 - T1 = {t2 - t1}")
print()
# Failure control: the paper's final step says T1 and T2 time the same interval.
print(f"control: same start event? {t_start == E[0]} -> the claim is refused")

Expected outcome

E: photon emitted    S ('0', '0')     S' ('0', '0')
A: M at S-time 0     S ('0', '1')     S' ('3/4', '5/4')
B: photon meets M    S ('5/8', '5/8') S' ('5/4', '5/4')

E and A simultaneous in M's frame? False
photon speed in M's frame: 1

T1 = D/(c+v) = 5/8, starting at E (S-time 0)
T2 = 1 = D/c? True, starting at M's event at S-time -3/8
the two start events differ by 3/8 in S-time; T2 - T1 = 3/8

control: same start event? False -> the claim is refused

Both times end at the same event, B, but they start at different ones: T1 at the emission E, T2 at an event on M’s worldline 3/8 earlier. Their difference is exactly the gap between the start events, so it is not a contradiction. The question is whether the derivation exposes an inconsistency in the mathematics, not whether relativity describes the world.

What the laboratory checked

The recorded run 20261113T232711Z-ca0a4a56 (evidence ) is the book’s result; the program above is a reader-sized demonstration of the same mathematics. Evidence: Verified historical kernel evidence (imported, not re-run) plus fresh exact computation. Status: historical_verified_correspondence_pending.

In the Lorentz model of the paper’s scenario, the paper’s two collision times T1 = D/(c + v) and T2 = D/c both end at the meeting event but start from different events, so their difference is not a contradiction; the three properties the paper names hold together in one Lorentz model.

Check System Recorded outcome
inputs Pinned inputs all inputs match their pins
inv02f-run2 Imported record (not re-run) 78 recorded files match the pinned manifest
compute Python (exact rational event ledger) exit 0, verdict different_start_events
compute-9-10 Python (exact rational event ledger) exit 0, verdict different_start_events

Failure controls: 1 of 1 rejected or refused as required (claim-same-start). Correspondence P03-CORR-01 (does the formal statement say what the claim means?): pending the author’s approval. Contract P03-HARAUX version 1, sha256 73c223c0f47584ea…; result.json sha256 2dbcfb9946758889….

To reproduce the laboratory’s run, build the math-env image from a clone of the repository (Getting Started ) and run:

img=$(docker image inspect math-env --format '{{.Id}}')
docker run --rm --network none -e UV_OFFLINE=1 -e MATHENV_IMAGE_ID=$img -v "${PWD}:/source:ro" -v "${PWD}/out:/out" math-env uv run --offline math-engine proof haraux --out /out/proof-haraux

This Docker form was tested for fermat, kempe and simpson, not separately for haraux. To re-verify the recorded run itself without Docker, math-engine proof --verify experiments/investigations/p03-haraux/runner-runs/2026-11-13-pinned (R-V03).

The runner re-verifies the imported MATH-INV-02F kernel record and recomputes the event ledger; it does not re-run the 84 kernel declarations. A full re-run is the separate procedure experiments/investigations/math-inv-02/run_02f.sh.

Mechanism and limitations

Documented: the claim contract experiments/investigations/p03-haraux/investigation.json (version 1) pins the statements, checks and controls before anything runs; the recorded run directory holds result.json, the runner’s notebook and a SHA256SUMS manifest of every file.

Observed: the program’s output above, and the recorded run 20261113T232711Z-ca0a4a56: Historical evidence imported and verified: the recorded run’s files match their pinned manifest byte for byte. No kernel was run in this execution; any fresh computation is reported separately. The correspondence awaits the author’s approval.

Not shown, as the laboratory’s own record states:

  • Anything about whether special relativity describes the physical world.
  • That our reconstruction reads the paper correctly: that is the correspondence, pending.
  • A fresh kernel execution: the kernel evidence is imported, verified byte for byte, not re-run. Re-running it means the full MATH-INV-02F procedure (run_02f.sh), which is outside the reader workflow.
  • A refutation of the author, who has informally conceded in a public comment but not retracted the paper.

Challenge this result

Copy this into the AI assistant of your choice, or work through it yourself. Check any finding against the recorded evidence.

Examine the reconstruction of Haraux's scenario (Qeios, version 2, DOI
10.32388/MFHAG1.2). Do T1 = D/(c + v) and T2 = D/c really start from different
events in the Lorentz model, and is the reconstruction a fair reading of the
paper? Keep separate what the imported kernel record shows and what the fresh
computation shows, and look for a reading of the paper under which the
argument works.

Chapter: https://programmer.ie/books/math/04-chapter/
Example: https://programmer.ie/examples/math/04-chapter/
Recorded run: P03-HARAUX 20261113T232711Z-ca0a4a56, status historical_verified_correspondence_pending; evidence: Verified historical kernel evidence (imported, not re-run) plus fresh exact computation
Source and evidence: https://github.com/ernanhughes/math/tree/main/experiments/investigations/p03-haraux/runner-runs/2026-11-13-pinned
Distinguish what the record shows from what you infer. If you cannot open a link, ask me for the material; do not invent it.

Canonical source , the recorded run and the program .

Understand this example

Copy this prompt into your AI tool. No code runs here.

Apply this example

Copy this prompt into your AI tool. No code runs here.