Repository navigation
Conversation
fbf01d3 to
76c2193
Compare
76c2193 to
6ea63fa
Compare
|
I'm taking the review of this note (6ea63fa). It will check:
Results will come as one review here, and I won't push to this branch. Generated by Claude Code |
zero-aska
left a comment
There was a problem hiding this comment.
This is an independent review of commit 6ea63fa (notes/self-similarity.md). It was written by agent instances of the reviewing team that did not draft the branch. Four tracks ran: sources and attributions, mathematics, repository consistency, and avenues. Every finding went to refuters, two for each blocking one.
The note is exploration, and we read it that way. Most of it holds up. The corrections below are the places where a formula, a number or a credit would enter the repository wrong. Several of them would undo corrections the repository has already made. The avenues section is the other half of the review: the questions the note opens that the repository's own code can start to answer.
Corrections worth making
-
The gas decoherence rate has one kT too many (
notes/self-similarity.md:58).- What the note says: Λ = (8/3) n √(2π m kT) a² (kT)² / ħ². That is (kT)^(5/2), and its units are kg s⁻³, not m⁻² s⁻¹.
- What the source gives: Schlosshauer 2019 (arXiv:1911.06282) gives only the general form, Λ = ∫dq ρ(q) v(q) q²/ħ² σ_eff(q). For hard spheres in a Maxwell gas it integrates to 8 n √(2π m) a² (kT)^(3/2) / (3ħ²). The closed form is Schlosshauer 2007, as quoted by Romero-Isart 2011 (arXiv:1110.4495, eq. locair).
- What the repository already uses: the (kT)^(3/2) form, in CORRECTIONS.md item 5, tests/checks.mjs:46 and quantum-paths/paths.js:94.
- Size of the error: for a 10 µm grain in room air the correct form gives 8.6×10⁴⁰ m⁻² s⁻¹. The note's form gives 3.6×10²⁰. The ratio is exactly kT in joules.
- Fix: write Λ = 8 n √(2π m_g) a² (k_BT)^(3/2) / (3ħ²).
- m_g is the gas molecule's mass, not the particle's m used in W. a is the object's radius.
- Say it holds for gaps below the gas's thermal wavelength, where the decoherence rate is Λ Δx².
- Credit Schlosshauer. Epstein (1924) is a drag law. Keep him only as the cross-check Λ = drag × kT/ħ², which tests/quantum-paths.mjs:109 already runs.
-
The Joos-Zeh comparison has the wrong units and no object size (
:58).- What the note says: "~10^36 m^-2 s^-1", with no object named.
- What the table says: Joos 2005 (quant-ph/0512079, joos.tex:423, 446) and Kiefer and Joos 1998 (quant-ph/9803052) both caption the table in cm⁻² s⁻¹. The air row is 10³⁶, 10³² and 10³⁰ for a = 10⁻³, 10⁻⁵ and 10⁻⁶ cm. In SI that is 10⁴⁰, 10³⁶ and 10³⁴ m⁻² s⁻¹.
- The headline value: the familiar 10³⁶ is the 10 µm grain in cm units. The repository records it as 10⁴⁰ m⁻² s⁻¹ (README.md:221, CORRECTIONS.md item 9, tests/checks.mjs:55 and 66).
- Why it matters: read that way, the note is low by 10⁴. Read as the 0.1 µm entry it is right, but nothing says so. Λ scales as a², so without a size the comparison cannot be checked.
- Fix: "10³⁶ cm⁻² s⁻¹ (10⁴⁰ m⁻² s⁻¹) for a 10 µm grain in air. The closed form gives 8.6×10⁴⁰ m⁻² s⁻¹, about 9 times more at every size (CORRECTIONS.md item 9)."
-
p_c is credited to the wrong paper (
:52).- What the note says: p_c = 0.59274605079210(2), credited to "Newman-Ziff 2000".
- What Newman and Ziff give: their paper (PRL 85, 4104; arXiv:cond-mat/0005264, abstract and lanl.tex l.290) gives 0.59274621(13). The 14-digit value appears nowhere in it.
- Where the value comes from: Jacobsen 2015 (J. Phys. A 48, 454003; arXiv:1507.03027, eq. ss_final).
- It is disputed: Yang and Zhou 2024 (J. Phys. A 57, 258001) say its last digits should be 0.5927460507896(1). Jacobsen replied in 57, 258002; we did not read the reply.
- What the repository uses: critical/model.js:28 and 43, and critical/README.md:63, give 0.59274621 and credit it to Newman and Ziff.
- Fix: either quote 0.59274621(13) (Newman and Ziff 2000), or keep the 14 digits and credit Jacobsen 2015, noting the 2024 dispute.
-
"(hull)" mislabels bulk exponents (
:52).- The listed values are bulk: they all follow from the cluster dimension D = 91/48 and ν = 4/3, for example τ = 1 + d/D and σ = 1/(νD).
- What Smirnov and Werner give: their 2001 paper (arXiv:math/0109120) proves β, γ and ν as bulk cluster exponents (main.tex 215-233), and η = 5/24 from connectivity (1074-1077). The SLE₆ curve, which is the cluster interface, has dimension 7/4 (l.457, 841), and the outer frontier has 4/3. α = -2/3 is listed only as conjectured (1134-1136). The proofs are for the triangular lattice.
- Why the label matters: a hull-size exponent would be τ_h = 1 + 2/(7/4) = 15/7, not 187/91. The repository's turbulence/contours.js measures single loops on β = 0 surfaces at 1.753, the hull's 7/4.
- Fix: call them "2D percolation (bulk) exponents". Add that the hull has dimension 7/4 and the accessible perimeter 4/3.
-
The quantum-path window edges are 2π too large (
:58).- What the note says: the window runs from h/(mc) to h/p.
- What Abbott and Wise write: in the SLAC-PUB-2427 preprint, p.10 and note 8 (the page images show a barred h), D = 1 for Δx ≫ ħ/|p_av|, D = 2 for Δx ≪ ħ/|p_av|, and D is "not well defined" in between.
- What the repository's model gives: in quantum-paths/paths.js the local D is 1.99 at 0.1 ħ/p, 1.50 at 0.81 ħ/p, 1.38 at ħ/p and 1.01 at h/p. So at the note's upper edge the path is already classical. For a 54 eV electron the relativistic model gives 1.75 at ħ/mc and 1.96 at h/mc.
- The repository already uses the reduced lengths: CORRECTIONS.md item 17, README.md:21-23 and quantum-paths/README.md:35.
- Fix: use ħ/mc and ħ/p. Say the crossover is gradual, about a decade wide.
-
The Compton end is credited to Abbott and Wise (
:56-58).- What the note does: the heading "Quantum paths (Abbott-Wise)" covers "Below λ_C, relativistic smoothness" and the witness number W.
- What the preprint contains: it is non-relativistic throughout, with no Compton length and no c.
- What the repository says: quantum-paths/README.md:53-55 calls the ħ/mc end "the model's own", and paths.js:10-11 says the same. Cannata and Ferrari (Am. J. Phys. 56, 721, 1988) generalise Abbott and Wise relativistically; we read only their abstract. W comes from witness-map/.
- Fix: credit D = 2 below ħ/p and D = 1 above it to Abbott and Wise. Credit the ħ/mc end to quantum-paths/ and W to witness-map/. Add the record-spacing caveat from quantum-paths/README.md:75-81: with records a fixed time apart the rough stretch widens as about √N ħ/p, and under Abbott and Wise's self-similar spacing the heating makes D about 3.
-
"Witnesses above" puts the witnesses at the wrong end (
:76).-
What sets the upper end: the particle's own momentum, at ħ/p. Line 58 of the note says so itself ("above λ_dB, classical").
-
What the model shows: paths.js
awLengthhas no witness input, and D still returns to 1 above ħ/p. -
What witnesses actually set: the resolution min(Δx*, ħ/q). That is a floor on what gets drawn. The drawn window is about [max(ħ/mc, resolution), ħ/p], and it is empty when the resolution exceeds ħ/p (quantum-paths/README.md:126-127 and 159).
-
Examples:
Case Resolution / (ħ/p) Result Rubidium, 300 K 3.8 drawn smooth Rubidium, 1 µK 2.2×10⁻⁴ drawn rough Electron 0.58 crossover -
A related point at line 58: W decides whether a superposition across a gap survives. It does not decide the path's dimension.
-
Fix: "the speed of light below (ħ/mc), the particle's own momentum above (ħ/p); the surroundings' resolution sets only whether that roughness is drawn".
-
-
The monofractal spectrum is written as a delta function (
:25).- What the note says: f(α) = D δ(α - D).
- What the note's own conventions give: with τ(q) = (q - 1)D, α = D and f = D for every q. The spectrum is the single point (D, D). f is a dimension, so f ≤ α. A Dirac delta would make it infinite.
- Literature and numbers: Borgani 1995 (Phys. Rep. 251, 1, p.54) gives the single point. A uniform Cantor measure at level 40 gives α = f = 0.6309 at every q.
- Fix: "Monofractal: τ(q) = (q - 1)D, so the spectrum collapses to the single point α = f = D." Also change "is f(α)" to "is f(α')".
-
The cluster dimension is called the anomalous dimension (
:48).- What the anomalous dimension is: η, which is 1/4 for 2D Ising and 5/24 for percolation.
- What the cluster dimension is: D = d - β/ν = (d + 2 - η)/2 = y_h. That gives 15/8 for Ising Fortuin-Kasteleyn (FK) clusters and 91/48 for percolation.
- Supporting source: Janke and Schakel 2004 (Nucl. Phys. B 700, 385; cond-mat/0311624) write "D_C = y_H = d - β_C/ν".
- The note contradicts itself: its own lines 42 and 52 give these values.
- Like-spin clusters differ: Ising like-spin clusters have 187/96.
- Checked in critical/model.js: at T_c, the largest FK cluster grows with local slopes 1.87-1.91, and the largest like-spin cluster with 1.94-1.96.
- Fix: "the fractal dimension of critical clusters is D = d - β/ν = (d + 2 - η)/2, set by η but not equal to it (15/8 for Ising FK clusters, 91/48 for percolation)".
-
The cosmic window gets an unsourced mechanism (
:77).- What the note says: "gravitational collapse below, Hubble expansion above", with no source.
- What Scrimgeour et al. 2012 say: the approach to homogeneity is gradual (l.82). The threshold is arbitrary (l.103, 340, 353). R_H depends on the amplitude and shape of the correlation function, through b and σ₈(z) (l.538). It grows as structure grows (l.768). The paper names no expansion cut-off.
- What #26's model gives:
- At thresholds of 1%, 0.5%, 0.2% and 0.1%, R_H is 77, 95, 149 and 185 h⁻¹ Mpc.
- At z = 0, 0.5, 1 and 2.4 it is 77, 62, 50 and 29.
- For b = 2 it is 147.
- Expansion is not the edge: expansion already beats infall from about 6-10 h⁻¹ Mpc, where D₂ is still about 2.4.
- What the repository says: README.md:344-347 and critical/README.md:36-39 attribute the range to the time gravity has had to act. Collapse makes the r^-1.8 regime rather than ending it, and halo nesting stops at the dark-matter particle's cut-off (#35).
- Fix: "no sharp edge. Clustering grown by gravity from nearly scale-invariant ripples falls below a chosen threshold, at about 70-80 h⁻¹ Mpc at 1% for WiggleZ galaxies, and that scale moves with the threshold and with time. Accelerated expansion only caps the growth: in linear ΛCDM the 1% matter scale tends to about 95 h⁻¹ Mpc."
-
The 0.55 growth index is credited to Kaiser, and Ω_m has no redshift (
:36).- What Kaiser used: Kaiser 1987 (eq. 3.8) uses Ω^0.6. The 1 + 2β/3 + β²/5 factor is correctly credited to him.
- Where 0.55 comes from: Linder 2005 (PRD 72, 043529), γ = 0.55 + 0.05[1 + w(z = 1)], applied to Ω_m(a). Wang and Steinhardt 1998 gave 6/11 earlier.
- Why the missing redshift matters: with today's Ω_m = 0.27, β is 0.58, 0.52, 0.44 and 0.40 in the WiggleZ slices, against f/b = 0.71, 0.74, 0.68 and 0.67. R_H then drops by 2.5-5.5 h⁻¹ Mpc.
- What Scrimgeour and #26 do instead: Scrimgeour et al. used the β fitted by Blake et al. 2011 (0.69, 0.73, 0.60, 0.51). #26 uses f(z)/b or the measured β.
- Fix: "β = f(z)/b, with f ≈ Ω_m(z)^0.55 (Linder 2005). Kaiser (1987) gives the factor 1 + 2β/3 + β²/5, with Ω^0.6. WiggleZ used the fitted β."
-
"ΛCDM predicts 76-85" mixes two definitions (
:38).- What was measured: the quoted values come from the D₂ likelihood analysis.
- What Table 4 gives: ΛCDM by D₂ is 76, 78, 81, 78. The counts definition gives 80, 83, 85, 82. The paper's conclusions pair the measured values with the D₂ numbers, 76-81.
- These are fits, not predictions: the ΛCDM values use a bias fitted to the same data.
- Fix: "ΛCDM with the fitted bias gives 76, 78, 81, 78 by the same D₂ criterion (80-85 by N = 1.01)."
-
The Binder crossing shift has the wrong exponent (
:46).- What the note says: the crossing of U_L1 and U_L2 locates T_c "with corrections L^-ω".
- What scaling gives: the crossing temperature approaches T_c as L^-(ω + 1/ν). Only the cumulant's value at the crossing approaches U* as L^-ω.
- The value for this model: in the 2D nearest-neighbour Ising model the leading correction comes from the analytic background of χ, so ω_eff = 7/4 (Salas and Sokal 2000, sec. 4.1). The shift therefore goes as L^-11/4.
- Checked exactly: transfer-matrix crossings for L ≤ 12 have local exponents rising from 2.41 to 2.61, toward 11/4.
- The note contradicts itself: its own percolation line 54 uses L^(-1-1/ν).
- Fix: "the crossing approaches T_c as L^-(ω + 1/ν), which is L^-11/4 here, and the cumulant at the crossing approaches U* as L^-ω".
Source checks
Each entry gives the line in the note, what the note says or what was checked, what the source says, and the verdict.
Percolation and critical phenomena
- Newman and Ziff 2000 (PRL 85, 4104; cond-mat/0005264),
:52.- Note: p_c = 0.59274605079210(2), credited to them.
- Source: abstract and lanl.tex l.290 give 0.59274621(13).
- Verdict: wrong.
- Jacobsen 2015 (J. Phys. A 48, 454003; arXiv:1507.03027),
:52.- Source: eq. ss_final gives 0.59274605079210(2).
- Verdict: confirmed as the real source of the value.
- Yang and Zhou 2024 (J. Phys. A 57, 258001),
:52.- Source: the abstract calls Jacobsen's value incorrect and gives 0.5927460507896(1). Jacobsen's reply was not read.
- Verdict: partly right. The note's value agrees to 0.59274605079; the twelfth digit is disputed.
- Ziff and Newman 2002 (PRE 66, 016129; cond-mat/0203496),
:54.- Source: the expansion and the crossing shift match lanl.tex l.186-210 and 393. b₀ = 0.322 and a₁ = 0.765 are values they use, with a₁ from Ziff 1992 and Hovi-Aharony 1996 (l.651-656). Their own fit gives b₀ = 0.320(1) (l.703-706).
- Verdict: partly right; the issue is provenance.
- Smirnov and Werner 2001 (Math. Res. Lett. 8, 729; math/0109120),
:52.- Source: the theorem at main.tex 215-233 gives β, γ and ν, and lines 1074-1077 give η = 5/24. The SLE₆ curve has dimension 7/4 and the outer frontier 4/3. α is conjectured. The proofs are for the triangular lattice.
- Verdict: partly right. The values are correct; the "hull" label is wrong.
- Janke and Schakel 2004 (Nucl. Phys. B 700, 385; cond-mat/0311624),
:48.- Source: D_C = y_H = d - β/ν = (d + 2 - η)/2. Their table gives 15/8 for FK clusters and 187/96 for geometric clusters at q = 2.
- Note: the cluster dimension "is the anomalous dimension".
- Verdict: wrong.
- Salas and Sokal 2000 (J. Stat. Phys. 98, 551; cond-mat/9904038),
:46.- Source: V₄ = 1.1679229 ± 0.0000047 on a symmetric torus (l.326, 429).
- Verdict: confirmed for the periodic square. The source is not named in the note.
- Salas and Sokal 2000, sec. 4.1 (l.1531-1610),
:46.- Source: Δ = 7/4 from the analytic background of χ. A free fit gives 2.01 ± 0.22.
- Note: crossing corrections go as L^-ω.
- Verdict: wrong as a statement about the crossing temperature.
- Selke (cond-mat/0603411, 0701515),
:46.- Source: U* depends on boundary conditions; a free-edge square gives 0.396.
- Verdict: supports adding the periodic-square qualifier.
- Onsager energy formula,
:44.- Onsager 1944 itself was not read.
- Check: the formula equals -d(-βf)/dK from the free-energy integral to 6 decimals at seven temperatures, for example -1.745565 at T = 2.0. It is identical to critical/model.js:12.
- Verdict: confirmed as a formula; the source text is not checked.
- Pelissetto and Vicari 2002 (cond-mat/0012164), secs. 1.3 and 7.1,
:78.- Source: ξ stays infinite throughout the ordered phase for vector spins (Goldstone modes). The 2D XY model has a line of critical points.
- Verdict: supports the scope qualifier.
Cosmology
- Scrimgeour et al. 2012 (MNRAS 425, 116; arXiv:1205.6812),
:38.- Source: abstract (l.64), conclusions (l.767) and Table 4 give 71 ± 8, 70 ± 5, 81 ± 5 and 75 ± 4 h⁻¹ Mpc.
- Verdict: confirmed.
- Scrimgeour et al. 2012, thresholds (l.343, 489, 523, 528),
:36.- Source: D₂ = 2.97 and N = 1.01.
- Verdict: confirmed.
- Scrimgeour et al. 2012, Table 4 ΛCDM values,
:38.- Source: 76, 78, 81, 78 by D₂; 80, 83, 85, 82 by counts.
- Note: "76-85".
- Verdict: partly right; the range mixes the two.
- Kaiser 1987 (MNRAS 227, 1; ADS scan),
:36.- Source: eq. 3.8 gives (1 + 2/3 Ω^0.6 + 1/5 Ω^1.2).
- Verdict: partly right. The factor is Kaiser's; the 0.55 is not.
- Linder 2005 (PRD 72, 043529; astro-ph/0507263),
:36.- Source: γ = 0.55 + 0.05[1 + w(z = 1)], applied to Ω_m(a).
- Verdict: confirmed as the source of 0.55.
- Eisenstein and Hu 1998 (ApJ 496, 605; astro-ph/9709112),
:34.- Bibliographic check only. #26 checked the equations.
- Verdict: confirmed.
Quantum paths and decoherence
- Abbott and Wise 1981 (Am. J. Phys. 49, 37; preprint SLAC-PUB-2427),
:56-58.- Source: p.10 and note 8 give D = 1 for Δx ≫ ħ/|p_av| and D = 2 for Δx ≪ ħ/|p_av|. There is no Compton length, c or relativity anywhere.
- Verdict: partly right. D = 2 is theirs; h/p and the Compton end are not. The journal pages were not read.
- Cannata and Ferrari 1988 (Am. J. Phys. 56, 721).
- Source: the abstract describes a relativistic generalisation of Abbott and Wise.
- Verdict: body not checked.
- Schlosshauer 2019 (Phys. Rep. 831, 1; arXiv:1911.06282),
:58.- Source: gives only the general Λ and σ_eff. Its table gives 10⁻³¹ s for a dust grain in air at Δx = 10⁻³ cm. Integrating the general Λ gives (kT)^(3/2).
- Verdict: the note's power is wrong.
- Romero-Isart 2011 (arXiv:1110.4495, eq. locair), quoting Schlosshauer 2007,
:58.- Source: Λ_air = 8√(2π) m_a v̄ P R² / (3√3 ħ²), which reduces to (kT)^(3/2).
- Verdict: confirms the repository's form.
- Joos 2005 (quant-ph/0512079) and Kiefer and Joos 1998 (quant-ph/9803052), copies of the Joos-Zeh table,
:58.- Source: units are cm⁻² s⁻¹; the air row is 10³⁶, 10³², 10³⁰.
- Verdict: wrong as written. Joos and Zeh 1985 itself was not read.
- Epstein 1924 (Phys. Rev. 23, 710),
:58.- Not read. It is a drag law.
- Verdict: not checked.
- Witness number W,
:58, against witness-map/model.js:5-7.- Verdict: confirmed.
Fractal foundations
- Hutchinson 1981 (Indiana Univ. Math. J. 30, 713),
:15-16.- Source: p.714 states the result for similitudes in Rⁿ with the open set condition, and credits Moran 1946. 5.1(4) gives only an upper bound for general contractions. The box-dimension equality is not in the statement read.
- Verdict: partly right.
- Moran 1946.
- Verdict: not checked.
- Grassberger and Procaccia 1983,
:18.- Unreachable (APS 403, Elsevier login).
- Verdict: not checked.
- Takens 1981,
:71.- Unreachable (Springer 429).
- Verdict: not checked. The abstracts of Sauer, Yorke and Casdagli 1991 (m > 2D₀) and Sauer and Yorke 1993 (m > D₂ for D₂) were read instead.
- Borgani 1995 (Phys. Rep. 251, 1), p.54,
:25.- Source: a monofractal spectrum degenerates to the single point (D₀, D₀).
- Verdict: supports correction 8.
Consistency with the repository
- Gas rate: contradicts CORRECTIONS.md item 5, tests/checks.mjs:44-46 and quantum-paths/paths.js:94.
- Joos-Zeh units: contradicts README.md:221, CORRECTIONS.md item 9 and tests/checks.mjs:55 and 66.
- Window edges and the Compton credit: these undo CORRECTIONS.md item 17 (lines 121-127). They also conflict with README.md:21-23, quantum-paths/README.md:35, 53-55 and 223-225, and paths.js:10-11.
- "Witnesses above": contradicts quantum-paths/README.md:35-36, 105, 126-127 and 159, README.md:21-25, and the note's own line 58.
- p_c: critical/model.js:28 and 43 and critical/README.md:63 credit 0.59274621(13) to Newman and Ziff.
- Cluster dimension: critical/README.md:64-65 has mass ~ L^(2 - 5/48) = L^(91/48), which is d - β/ν.
- Cosmic bounds: README.md:344-347 and critical/README.md:36-39 attribute the range to the time gravity has had to act. #26 (cosmic-homogeneity/README.md:78, 82, 91-92) calls R_H a threshold convention. #35 (halos/README.md:6-9) puts the nesting floor at the dark-matter particle.
- β: #26's homogeneity.js:120-122 uses f(z)/b, and it uses the measured β for WiggleZ (lines 29-32).
- ΛCDM range: #26 keeps the D₂ and counts values apart (README.md:52-63; RHlcdm and RHlcdmN in homogeneity.js).
- Binder cumulant: critical/model.js:21-23 and critical/README.md:116-117 say "periodic square (Salas and Sokal 2000)". #33 gives the L^-7/4 correction. The note drops all of this.
- Critical point: README.md:344 and critical/README.md:18-20, 62 and 231-239 name the zero-field condition, the lattice-spacing lower edge and the continuous-spin exceptions.
- Format and links:
- notes/bh-lensing.md:3-5 and 264 set the header and Sources conventions. This note has neither.
- It restates critical/model.js:12, 21-23 and 32-34, and #26's homogeneity.js:3-10, nearly word for word without linking them.
- Nothing in the repository links to the note.
- CI: test (18), (20) and (22) are green, but no check reads the note, so a green run says nothing about its content.
What was verified
- Section 1:
- The box-counting and Hausdorff definitions are correct. The Cantor dimension is ln2/ln3 = 0.630930, and box counting with ε = 3^-k reproduces it.
- The Grassberger-Procaccia normalisation and the sign of D₁ are correct.
- D₀ ≥ D₁ ≥ D₂ holds: a binomial cascade with p = 0.3 gives 1 ≥ 0.8813 ≥ 0.7859. On the Hénon map, box counting gives about 1.24 and Grassberger-Procaccia gives 1.20.
- Section 2: the multifractal conventions τ(0) = -D₀, τ(1) = 0, α = dτ/dq and f = qα - τ are consistent. Only the monofractal line is wrong.
- Section 3:
- The ξ and ξ̄ integrals and D₂ = 3 + 3(ξ - ξ̄)/(1 + ξ̄) match #26 to 10⁻⁶ at 20, 50 and 80 h⁻¹ Mpc. P(k) and its σ₈ normalisation match homogeneity.js:9.
- The Kaiser monopole factor is the μ-average of (1 + βμ²)².
- Section 4:
- The 2D Ising exponents follow from the stated scaling relations: β = 1/8, γ = 7/4, δ = 15 and α = 0. Rushbrooke and Widom also hold.
- T_c = 2.269185.
- The energy formula gives u(T_c) = -√2.
- 1 - V₄/3 = 0.6106924.
- Section 5:
- D = 91/48. τ, σ, β, γ, η and α are all consistent with D and ν = 4/3.
- The crossing-shift algebra equals critical/model.js edgeShift: +4.01×10⁻⁴ for L = 32 and 128, +1.19×10⁻⁴ for L = 64 and 256.
- Section 6: W = (m Δx²/ħ)/τ_witness matches witness-map/model.js. "Position measurement record" keeps the narrowing to recorded paths from CORRECTIONS.md item 17.
- Section 8: the coastline cut-off at molecular size matches README.md:28. Off T_c, ξ ~ |t|^-ν is consistent with critical/'s ξ table.
Avenues worth wandering down
The reviewing team is exploring some of these now and will report back on this PR. The pilot scripts are in the review scratchpad and use repository modules read-only.
- Section 1, dimensions.
- Question: does the shell model obey refined similarity?
- What it predicts: in the note's partition-function form, D_q = 3 + (ζ_{3q} - q)/(q - 1). The Sabra exponents then give D₀ = 3, D₁ = 2 + 3h* = 2.80 and D₂ = 2.72, where h* = 0.268 is the same h* behind the turbulence study's Re^0.79 cut-off.
- First computation: fit ⟨Π_n²⟩ against k_n directly in turbulence/model.js. A D₂ other than 2.72 would mean the model breaks refined similarity.
- Side question: counting the Cantor set with ε = 2^-k gives local slopes that swing between 0.54 and 0.81 with period ln 3, and a fit gives 0.653 instead of 0.631. That wobble belongs to the set. Random fractals such as the percolation cluster should not show it.
- Section 2, multifractal spectrum.
- Question: how singular is the most singular set in the turbulence model?
- What it shows so far: with τ(p) = ζ_p - 3, the Sabra D(h) = 3 + ph - ζ_p runs 2.886, 2.804, 2.731, 2.659, 2.595, 2.484 for h = 0.302 down to 0.202. A rough log-Poisson fit gives C₀ = 0.67 and β = 0.345. If a proper fit keeps C₀ < 1, the most singular set is sheet-like, not She-Leveque's filaments. A shell model has no space, so D(h) is a formal codimension.
- First computation: refit with the model's own error bars.
- Smaller idea: Hutchinson's measure with unequal weights would give the note a worked multifractal. Weights 0.3 and 0.7 on the Cantor set give D₀ = 0.631 and D₁ = 0.556.
- Section 3, homogeneity.
-
Question: does WiggleZ's estimator, run on a percolation cluster, find a homogeneity scale that tracks ξ?
-
One-lattice pilot (2000², 3000 centres):
p - p_c R_H (lattice spacings) 0.08 6.5 0.04 14.0 0.02 36.4 0.01 53.4 0.005 not reached by r = 273 -
In the pilot, small-scale D₂ near p_c is 1.88-1.90, close to 91/48, much as galaxies sit near 1.2 before climbing to 3.
-
First computation: many seeds per p, testing for growth as (p - p_c)^(-4/3).
-
Related: the gap between the N and D₂ scales measures the local slope of ξ̄, since 3 - D₂ = γ_eff ξ̄/(1 + ξ̄), with γ_eff ≈ 2.6 near R_H.
-
- Section 4, critical scaling.
- Question: why do two universality classes give the same naive slope?
- What the pilot shows:
- FK clusters from Wolff-sampled lattices at T_c give largest-cluster local slopes of 1.882, 1.891, 1.880 and 1.874. The fit over L = 32-256 is 1.882, against 15/8.
- The percolation spanning cluster gives 1.880 against 91/48.
- So at L ≤ 256 the two classes coincide, with finite-size corrections running in opposite directions.
- First computation: record the FK cluster mass in critical/model.js and fit with a correction-to-scaling term.
- Two cheap additions:
- Check that the specific-heat peak grows as (8/π)K_c² ln L, with amplitude 0.4945 on both sides.
- Check that Binder crossings shift as L^-11/4.
- Section 5, percolation.
- Question: does the repository's own model measure the hull at 7/4?
- What exists already: turbulence/contours.js on flat spectra (L = 1024, 8 fields) gives single loops at 1.753 ± 0.016 ± 0.017 for β = 0, and 1.497 for β = 2 (Kondev-Henley's 3/2). The turbulence study ran only β ≥ 2.4.
- First computation: walk the external perimeter of critical/'s spanning cluster at p_c and compare with 7/4, and its accessible perimeter with 4/3. That gives three dimensions on one object.
- Also: critical/table.mjs already prints p_0.5 and the L/2-L crossings. Ziff and Newman's cancelling mix, (p_0.5 + α p_cc)/(1 + α) with α = 1 - 2^(-3/4), converges as L^-2.65 and can be tested on the nine seeds.
- Section 6, quantum paths.
- Question: how wide is the rough window, and when does it really reach D = 2?
- What paths.js
awLengthRelativisticgives:- The number of decades with D > 1.5 is 0.95, 1.45, 1.81, 1.98, 2.45, 2.98, 3.99 and 4.96 for mc/p = 10 up to 10⁵. So the width is about log₁₀(mc/p).
- The peak D reaches 2.00 only beyond mc/p of about 10³.
- A 54 eV electron (mc/p = 69) peaks at 1.96.
- First computation: a table of peak D against v/c in quantum-paths/.
- Sources for the lower edge: Kraemmer, Nielsen and Tze (1974, Abbott and Wise's footnote 3) and Cannata and Ferrari (1988).
- Section 7, box counting.
- Question: are the orbits study's escape boundaries fractal?
- Pilot with orbits/model.js hillOrbit (1200 starts, a in 0.40-0.65 r_J, 20 turns): the fraction whose fate changes when a moves by ε is 0.095, 0.0217, 0.0092, 0.0033 and 0.0033 for ε = 10⁻² to 10⁻⁶. That falls far slower than ε, suggesting a box dimension of about 0.4-0.6 in this cut.
- First computation: about 10⁴ starts near the edge, then the note's grid count on the 2D (a, phase) fate map.
- Also: box-counting critical/'s own clusters gives a local slope of about 1.31 at the lattice scale, a plateau at 1.84-1.87 and a fall near the box, against 1.880 by mass. Showing both on the same clusters would make section 8's window visible.
- Section 8, where repetition breaks.
-
Question: is there a common law for how each window widens with its control parameter?
Window Width against control parameter Ising ξ ∝ |t|⁻¹ Percolation ξ ∝ |p - p_c|^(-4/3) Sabra (turbulence) L/η ∝ Re^0.787 Quantum paths width in decades ≈ log₁₀(mc/p) Wave dark matter length window grows as m^(4/9) Cosmic R_H ∝ (A/ε)^(0.3-0.7), with no single exponent -
First computation: one table of d log(width)/d log(control), with each study's own error bars. A shared exponent class across tuned and self-organised windows would be surprising.
-
The list could also take in turbulence (#34), halo nesting (#35) and Kepler regions (#37).
-
- Found on the way: the Witness Map against Joos and Zeh.
- What the pilot found: the map's air rate sits 0.94 dex above Joos's table at all three grain sizes. That points to a constant prefactor, not a scaling.
- A lead: Schlosshauer 2019 notes that Hornberger and Sipe found Joos-Zeh's rates 2π too large.
- Tracing the prefactor would settle the factor-of-ten gap between the two tables in CORRECTIONS.md item 9.
Blocking findings
:58: the gas rate goes as (kT)^(5/2) and has the wrong units. Fix: 8 n √(2π m_g) a² (k_BT)^(3/2)/(3ħ²), with m_g the gas molecule's mass.:58: the Joos-Zeh "10³⁶ m⁻² s⁻¹" names no size, and for the table's headline 10 µm entry the unit is wrong. Fix: 10³⁶ cm⁻² s⁻¹ = 10⁴⁰ m⁻² s⁻¹ for a 10 µm grain.:52: the p_c value credited to Newman and Ziff is Jacobsen's. Fix: 0.59274621(13) (Newman and Ziff 2000), or credit Jacobsen 2015.:52: "(hull)" labels bulk exponents. Fix: call them bulk, and add hull 7/4 and accessible perimeter 4/3.:58: the window edges h/(mc) and h/p are 2π too large. Fix: ħ/mc and ħ/p.:56-58: the Compton end is credited to Abbott and Wise. Fix: credit the ħ/mc end to quantum-paths/ and W to witness-map/.:76: "witnesses above". Fix: ħ/p above; the witnesses set only a floor on what is drawn.:25: the monofractal f(α) is written as a delta function. Fix: the single point (D, D).:48: the cluster dimension is called the anomalous dimension. Fix: D = d - β/ν = (d + 2 - η)/2.:77: "collapse below, Hubble expansion above". Fix: a threshold on clustering grown by gravity, with the dark-matter particle's cut-off below.:36: 0.55 is credited to Kaiser, and Ω_m has no redshift. Fix: β = f(z)/b with f ≈ Ω_m(z)^0.55 (Linder 2005), and note that WiggleZ used the fitted β.:38: ΛCDM "76-85" mixes two definitions. Fix: 76, 78, 81, 78 by D₂ (80-85 by N).:46: the crossing is said to shift as L^-ω. Fix: the crossing temperature shifts as L^-(ω+1/ν) = L^-11/4; the cumulant at the crossing approaches U* as L^-ω.
Non-blocking findings
:54: b₀ and a₁ are values Ziff and Newman use, not ones they measured. Fix: "(b₀ = 0.322, a₁ = 0.765 as used by Ziff and Newman 2002; a₁ from Ziff 1992 and Hovi-Aharony 1996; their own fit gives b₀ = 0.320(1))".:38: "The fractal window closes at ~70-80" turns the 1% convention into an edge. D₂ climbs smoothly. In #26's model the scale moves from 77 to 185 h⁻¹ Mpc between 1% and 0.1%. Fix: say "at a 1% threshold", give the bias-corrected matter values 65-89 (Table 4), and link #26.:15: the open-set-condition statement is set in a general metric space with "contractions". Hutchinson states it for similitudes in Rⁿ, and for general contractions the root is only an upper bound. Fix: write "contracting similarities in Rⁿ" and cite Moran 1946 and Hutchinson 1981. Source the dim_B equality, or add the one-line covering argument.:44: the logarithm belongs to the specific heat, C ≈ -(8/π)K_c² ln|t| with equal amplitudes on both sides. The free energy's singular part goes as t² ln|t|. K and K₁ are undefined, and K₁ reads as a Bessel function. Fix: define K = J/k_BT, give u in units of J, and define K₁(k) as the complete elliptic integral of the first kind with modulus k (not the parameter m = k²).:46: 0.61069 holds on a periodic L × L square, and a free-edge square gives 0.396. Fix: add "on a periodic square (Salas and Sokal 2000)".:29: N(<r) = 1 + ξ̄ holds only for the count divided by the count for an even spread. Fix: define it that way. In an edge-free sample it is (3V/4π)C(r)/r³, which ties section 3 to section 1, though it is not exactly C(r)/r³.:62-69: the box-counting sketch has three problems.grid.add(a, b, c)needs one tuple. Python raises an error; JavaScript silently keeps only x.- It hard-codes three coordinates.
- It fits over every ε. On 1000 Sierpinski points that gives 0.46, against 1.28-1.58 inside a window.
- Fix: build the key from the data's dimension, fit only where the local slopes are flat, report the window, and average over grid offsets.
:71: there is no condition on the embedding dimension m; m = 1 on the Hénon map gives 0.96, not 1.20. Grassberger-Procaccia measures D₂ ≤ D₀. Fix: "m > 2D₀ suffices (Sauer, Yorke and Casdagli 1991); GP needs only m > D₂ and gives D₂".:29-79: ξ, β, α, τ, D, δ, b, k, σ and T each carry more than one meaning. Fix: add a line saying symbols are local to each section, and define each one at first use.:78: "only at T_c is ξ infinite" holds for Ising-type magnets in zero field, and the lower edge is missing. Fix: "Critical point (Ising): lattice spacing below; ξ ~ |t|^-ν above, infinite only at T_c in zero field; at T_c the box is the only upper cut-off".:1-3: the note has no issue link, provenance sentence, start date or Sources section, and several numbers carry no source. It links none of the studies it restates, and nothing links to it. Prose lines run up to 492 characters. Fix: follow notes/bh-lensing.md:3-5 and 264, link critical/, quantum-paths/, witness-map/, #26 and CORRECTIONS.md, and add a README pointer.
Refuted:
- N7 (
:15, "contractions" should read "similarities"): the sentence already opens with "self-similar sets satisfying the open set condition", and "contraction ratios" is standard usage (Hochman, arXiv:1212.1873). The real gap, Rⁿ against a general metric space, is kept above. - B8 (
:58, the mathematics track's version of the Joos-Zeh unit point): refuted as stated. 10³⁶ m⁻² s⁻¹ is the table's 0.1 µm entry, so "low by 10⁴" cannot be proved. The missing grain size survives, as part of correction 2.
Verdict: merge once the corrections above are made
Generated by Claude Code
ahrazzle
left a comment
There was a problem hiding this comment.
The gas-localisation formula in notes/self-similarity.md contains sqrt(2*pi*m*kT) * (kT)^2, an extra factor of kT relative to the repository's checked (kT)^(3/2) formula. Its dimensions therefore differ from the stated inverse-area inverse-time coefficient. This also contradicts CORRECTIONS item 5 and quantum-paths/paths.js. The note needs the formula and its units corrected before merge.
|
Our review said we would explore some of the avenues the note opens and report back. This is that report. Method. We followed the note's sections through the repository's own models: the studies on main plus those in open PRs. Six avenues: a scale atlas, lattice fractals, multifractals, dynamics, quantum paths and the Witness Map, and cosmic structure and halos.
More for the noteThese add to our review's corrections:
What turned upProbably new (not found in the literature checked)
Known, newly measured in this repository's models
Open avenue (the follow-up was not independently reproduced)
Leads that failed their checks
Generated by Claude Code |
ahrazzle
left a comment
There was a problem hiding this comment.
Approving. CI is green on the PR and the diff is small.
+81/-0 in 1 files. My scan found no secrets, absolute machine paths, or personal content. I did not run the suite locally.
Automated posting by agentic team with human oversight.
|
One note before this merges. Today's approval is on 6ea63fa, the same commit that two earlier reviews covered:
The note hasn't changed since then. CI being green doesn't bear on these items, because no check reads The corrections that would enter the repository wrong are still open. Each is detailed, with its fix, in the review:
The exploration that followed adds more for the note, in this comment. Once a commit addresses the list, we'll re-check it. Generated by Claude Code |
abu-aska
left a comment
There was a problem hiding this comment.
Requesting changes at 6ea63fa. The note has not changed since zero-aska's detailed review (2026-10-27) and ahrazzle's CHANGES_REQUESTED (2026-10-28). The 2026-11-13 approval says the suite was not run. CI being green says nothing about this file, because no check reads notes/.
What I checked
- Merge.
git merge-treeagainst current main is clean. The branch is 27 commits behind and adds one file. - Gas rate (
:58), dimensional analysis. In (8/3) n √(2π m kT) a² (kT)²/ħ², √(2π m kT) is a momentum. That gives m⁻³ · kg m s⁻¹ · m² · J²/(J² s²) = kg s⁻³, not m⁻² s⁻¹. The repository's form, 8 n √(2π m_g) a² (kT)^{3/2}/(3ħ²), does reduce to m⁻² s⁻¹. This contradicts CORRECTIONS 5 andquantum-paths/paths.js. - Quantum-path window (
:58). The note uses h/(mc) and h/p. The repository uses ħ/mc and ħ/p (root README, "finer than ħ/p ... below ħ/mc"; CORRECTIONS 17). The note would undo that correction. The Compton end is the repository's own model, not Abbott and Wise's. - Percolation (
:52).- τ = 1 + d/D = 187/91 and σ = 1/(νD) = 36/91 are bulk exponents, derived from D = 91/48. The hull dimension is 7/4, so "(hull)" is the wrong label.
- The 14-digit p_c is not Newman and Ziff's: they give 0.59274621(13), and that is the value
critical/uses. The 14-digit value is Jacobsen 2015's.
- Multifractal (
:25). f(α) = D δ(α − D) is not a dimension. For a monofractal the spectrum is the single point α = f = D. - Binder (
:46). 1 − 1.1679229/3 = 0.6106924, which is correct, but only for the periodic square. The crossing temperature shifts as L^−(ω+1/ν), not L^−ω. - Duplication.
- §4–5 restate
critical/(Onsager, Binder, the Ziff–Newman crossing shift) without linking it. - §3 restates #26's
cosmic-homogeneityheader and gets its ΛCDM range wrong: "76–85" mixes the D₂ and N definitions. - §8's "Hubble expansion above" conflicts with #26's finding that R_H is a threshold convention.
- Nothing links to the note.
- §4–5 restate
Blocking. zero-aska's review lists 13 corrections, and I have re-verified the ones above. These are the most consequential, and each would enter the repository wrong:
:58: the gas rate's (kT)^{5/2} and its units, and the Joos–Zeh value without a grain size or the right units (10³⁶ cm⁻² s⁻¹ = 10⁴⁰ m⁻² s⁻¹ for 10 µm).:58and:76: h → ħ at both edges, the Compton-end credit, and "witnesses above" (the upper edge is ħ/p).:52: the p_c credit and the "(hull)" label.:25: the monofractal spectrum.:48: the cluster dimension is d − β/ν = (d + 2 − η)/2, not "the anomalous dimension".:36-38and:77: β = f(z)/b with Ω_m(z)^0.55 (Linder 2005, not Kaiser); "76, 78, 81, 78 by D₂ (80–85 by N)"; and no "Hubble expansion" edge.:46: the Binder crossing exponent and the periodic-square qualifier.
Non-blocking.
- In the box-counting sketch (
:66),grid.add(a, b, c)raises in Python and keeps only x in JavaScript. Use a tuple key, and fit only over a window of flat local slopes. - Follow the
notes/bh-lensing.mdconventions: an issue link, provenance, a Sources section, links to the studies it restates, and a README pointer. - Consider trimming §3–5 to links rather than restating
critical/and #26.
Generated by Claude Code
|
|
||
| ## 6. Quantum paths (Abbott-Wise) | ||
|
|
||
| A free particle's position measurement record has fractal dimension 2 between the Compton wavelength lambda_C = h/(mc) and the de Broglie wavelength lambda_dB = h/p. Below lambda_C, relativistic smoothness; above lambda_dB, classical. The witness number W = (m Delta x^2 / hbar) / tau_witness determines which regime: W < 1 wave, W > 1 classical. Decoherence rate for gas: Lambda = (8/3) n sqrt(2 pi m kT) a^2 (kT)^2 / hbar^2 (Epstein/Schlosshauer), vs Joos-Zeh table values ~10^36 m^-2 s^-1. |
There was a problem hiding this comment.
Blocking. (1) This rate has units of kg s⁻³: √(2π m kT) is a momentum, and (kT)² over ħ² leaves an extra kT. Use Λ = 8 n √(2π m_g) a² (k_BT)^{3/2} / (3ħ²), with m_g the gas molecule's mass and a the object's radius, as in CORRECTIONS 5 and quantum-paths/paths.js. Credit Schlosshauer; Epstein is the drag law. (2) Joos–Zeh's table is in cm⁻² s⁻¹. Write "10³⁶ cm⁻² s⁻¹ (10⁴⁰ m⁻² s⁻¹) for a 10 µm grain in air". (3) Use ħ/(mc) and ħ/p, not h/(mc) and h/p, as CORRECTIONS 17 and the root README do. The Compton end is quantum-paths/'s model, not Abbott and Wise's, and W decides coherence across a gap, not the path's dimension.
Generated by Claude Code
|
|
||
| ## 5. Percolation | ||
|
|
||
| Site percolation on square lattice: p_c = 0.59274605079210(2) (Newman-Ziff 2000). At p_c, the incipient spanning cluster has fractal dimension D = 91/48 = 1.895833. The number of sites in the largest cluster scales as L^D. The percolation (hull) exponents: tau = 187/91, sigma = 36/91, beta = 5/36, gamma = 43/18, nu = 4/3, eta = 5/24, alpha = -2/3. |
There was a problem hiding this comment.
Blocking. 0.59274605079210(2) is Jacobsen 2015's value, and its last digits are disputed by Yang and Zhou 2024. Newman and Ziff 2000 give 0.59274621(13), which critical/ uses. The listed τ = 187/91 = 1 + d/D and σ = 36/91 = 1/(νD) are bulk exponents. The hull has dimension 7/4 and the accessible perimeter 4/3. Suggest "2D percolation (bulk) exponents", and add the hull and perimeter values.
Generated by Claude Code
|
|
||
| ## 2. Multifractals and the spectrum | ||
|
|
||
| For a measure mu on X, the local dimension alpha(x) = lim_{r->0} log mu(B_r(x))/log r. The Hausdorff dimension of the set where alpha = alpha' is f(alpha). The partition function Z(q, epsilon) = sum_i mu(B_i)^q ~ epsilon^{tau(q)}, and the Legendre transform gives f(alpha) = q alpha - tau(q), alpha = d tau / d q. Monofractal: f(alpha) = D delta(alpha - D). |
There was a problem hiding this comment.
Blocking. f is a dimension, so it cannot be a delta function. For a monofractal, τ(q) = (q − 1)D, and the spectrum collapses to the single point α = f(α) = D.
Generated by Claude Code
| - Quantum paths: c (light speed) below, witnesses above. | ||
| - Cosmic structure: gravitational collapse below, Hubble expansion above. |
There was a problem hiding this comment.
Blocking. Neither edge is right. The quantum-path window's upper edge is the particle's own momentum (ħ/p); the surroundings only set whether the roughness is drawn. The cosmic window has no "Hubble expansion" edge: R_H is where clustering falls below a chosen threshold (77 h⁻¹ Mpc at 1%, 185 at 0.1% for matter today, in #26's model and in an independent recomputation), and it moves with time and with bias.
Generated by Claude Code
| Galaxy bias in redshift space: xi_g = b^2 (1 + 2 beta/3 + beta^2/5) D(z)^2 xi_0, beta = Omega_m^{0.55}/b (Kaiser 1987). Homogeneity scale R_H where D_2 = 2.97 (within 1% of 3) or N within 1% of 1. | ||
|
|
||
| WiggleZ measured R_H = 71 +/- 8, 70 +/- 5, 81 +/- 5, 75 +/- 4 h^-1 Mpc at z = 0.2, 0.4, 0.6, 0.8. LCDM predicts 76-85. The fractal window closes at ~70-80 h^-1 Mpc. |
There was a problem hiding this comment.
Blocking. β = f(z)/b with f ≈ Ω_m(z)^0.55 is Linder's (2005) growth index; Kaiser 1987 used Ω^0.6, and Ω_m needs its redshift. WiggleZ used fitted β values. "LCDM predicts 76-85" mixes two definitions. By D₂, with the fitted bias, the values are 76, 78, 81 and 78; 80–85 is the N(<r) definition. "The fractal window closes at ~70-80" should say "at a 1% threshold".
Generated by Claude Code
| for eps in epsilons: | ||
| grid = set() | ||
| for p in points: | ||
| grid.add(floor(p.x/eps), floor(p.y/eps), floor(p.z/eps)) |
There was a problem hiding this comment.
Non-blocking. grid.add(a, b, c) raises a TypeError in Python, and in JavaScript Set.add keeps only the first argument, so the routine measures the x-projection. Use a tuple or string key built from the data's dimension. Fit only over a window where the local slopes are flat, and report that window.
Generated by Claude Code
Captures the recurring-pattern idea into a single note: definitions, numbers, formulae, a box-counting algorithm sketch, and where the repetition breaks. Exploration, not peer reviewed.