Lineage of TraTT

Where TraTT came from, which script produced which notebook entry, which claims stand, which were withdrawn, and how to rerun the main computations. Written 2026-10-23 for the repository audit (NaNIsaac BACKLOG A2) from git history, notes/lab_notebook.md, notes/paper_draft.md and Isaac's project files. It adds no new result and does not reinterpret any; where a status is quoted it is quoted from the notebook or the draft.

The research notes (notes/) belong to the TraTT research lane and are not changed here. Entries 68-88 and notes/paper_draft.md are on the branch research/paper-viability (PR #16, draft), not yet on main; links below say so.

1. Earlier work and names

Name	What it was	Where the record is	Relation to TraTT
FTT (Floquet-Takagi-Taupin)	Isaac's first line on X-ray diffraction from a crystal modulated by a bulk acoustic wave: Floquet channels of the two-beam Takagi-Taupin (TT) equations. Manuscript drafts "Floquet-Takagi-Taupin" v1 and v2, slides "Floquet-Takagi-Taupin" and "DynamicFTT"	Isaac's Drive project folder 01_Projects/X-AOM/ (05_Paper/, 04_Slides/); not in any repository	same physics question. TraTT's first design document was titled "DynamicFTT" (DynamicFTT_Research_Handoff.md, added in 1216ff8, 2026-09-30, removed from the tree in 7438afe, 2026-10-15; still in history)
calculate	first code of the FTT line (2026-01-15, 0bff8d2): FTTSolver with Euler and RK45 TT integrators for a travelling-wave phase, and a notebook of Bessel sideband weights and satellite angles	NaNIsaac/calculate (private); its working copy is in the same Drive folder, 02_Src/calculate	one function copied: explore/calculate_rk45.py (b91b15b, 2026-10-16) keeps the adaptive RK45 integrator. TraTT's solvers in src/tratt/ were written separately and are not derived from it
FTT package (src/ftt/)	a later FTT code base described in Isaac's notes of 2026-08-04: tt.py (stroboscopic RK4 TT solver, Laue), model.py (travelling drive, parameters S, q, w), floquet.py (channel projection), metrics.py, experiment.py, tdtt.py (time-derivative channel-coupled engine), transient/ (Landau-Zener overshoot, realization map)	not found on this PC or in any NaNIsaac repository	TraTT has modules with the same roles (core.py, drive.py, floquet.py, observe.py); whether code moved cannot be checked without the source
AW_TT_EXP	an experiment-feasibility calculator (spec_calc.py): from X-ray energy, acoustic frequency, sound speed, Bragg angle, rocking width, crystal thickness, pixel size and distance, decides whether the ±1 acoustic satellites are separable in angle or on the detector	described in the same notes; source not found	no counterpart in TraTT; the realisability tables of Entries 20, 73 and 85 answer a related question with verified constants
TrTT	the name used for TraTT's KPS 2026 Fall poster preparation (TrTT_presentation_handbook, TrTT_last_minute, in 01_Projects/Conferences/2026_KPS_Fall/)	Drive	the same project; the handbook names the repository NaNIsaac/TraTT and the package tratt

Order in time, from git and file dates (Drive file dates are copy dates and only bound the authoring date from above): FTT manuscript drafts and slides by early 2026; calculate 2026-01-15; FTT package by 2026-08-04; TraTT initial commit e761142 on 2026-08-28 (DESIGN.md, core.py, drive.py, tests); transient-formation pipeline and notebook Entries 0-9 on 2026-09-30; poster at KPS 2026 Fall on 2026-10-22; paper-viability work (Entries 68-88) on 2026-10-22/23.

calculate's README said "Merged into TraTT"; that is being corrected in NaNIsaac/calculate#1. Only the one integrator above was copied.

2. Scripts and notebook entries

Scripts are in explore/nb/ unless a path is given; explore/make_*.py draw the poster and notebook figures. Run from the repository root with uv run python <script> [args]; the arguments are the ones written in the entry or in the script's docstring. "(PR #16)" marks scripts that are only on research/paper-viability.

Entry	Topic	Script(s)
0	starting point	—
1	formation order, n = 0 two-layer law	e1_order.py
2	+12 % excess is the static tail	e2_tail.py
3	peak shift is kinematic (negative)	e3_shift.py, e3b_shift_scan.py
4	satellite width narrowing	e4_width.py
5	finite rise time, accumulated coupling	e5_rise.py
6	units, temporal Pendellösung period	e6_units.py
7	prior art	—
8	transient ±n asymmetry	e8_asym.py
9	Bragg geometry, t_f ~ 1/α_n	e9_bragg.py
10	poster summary figure	explore/make_poster_figure.py
11	absorption (estimate only)	e11_absorb.py
12	claims and evidence (first table)	—
13	in-situ S calibration	e13_calib.py
14	three-mode, two-layer model	e14_threemode.py
15	ramp through the Bessel ladder	e15_cancel.py
16	validity of the filled-depth law	e16_validity.py
17	Bragg satellite width	e17_bragg_width.py
18	overshoot table without a fit	explore/make_feature_figure.py
19	random conditions	e19_random.py
20	experiment mapping, X0h constants	e20_units_x0h.py
21	prior art, second pass	—
22	time-resolved rocking curve picture	explore/make_map_figure.py
23	dark point, inward fringes	e23_dark.py
24	one front, three satellites	explore/make_ladder_map.py
25	Bessel zero, slow clock stops	e25_zero.py
26	regime map	explore/make_regime_map.py
27	claims table, additions 20-26	—
28	transient-only satellite	e28_transient_only.py, memory_drive.py
29	Polikarpov & Skadorov read	—
30	formation remembers the path	e30_history.py, explore/make_memory_figure.py
31	brightest order switches; grid-rule bug	e31_switch.py
32	two co-propagating waves	e32_two_waves.py
33	claims table, additions 28-32	—
34	is the transient-only satellite real	—
35	its robustness and failure boundary	e35_robust.py, nonideal.py, explore/make_robust_figure.py
36	up vs down (mostly negative)	e36_updown.py, explore/make_updown_figure.py
37	two acoustic domains	e37_domains.py, explore/make_domain_figure.py
38	chirped waves, Landau-Zener plateau	e38_chirp.py, qmod.py
39	acoustic domain walls	e39_walls.py
40	automated anomaly search	e40_search.py, explore/make_search_figure.py
41	validation and conservative summary of 38-40	e41_validate.py
42	exact Floquet solver (src/tratt/floquet.py)	e46_bands.py
43	regime map, micromotion	e43_phase.py, explore/make_phase_figure.py
44	hierarchy of Floquet paths	e44_mechanism.py, e44b_sidebands.py
45	Bessel-zero resurrection	e45_zero.py, e45b_shifted_zero.py, explore/make_zero_figure.py
46	Floquet bands, gap folding	e46_bands.py, e48_trajectories.py, explore/make_floquet_figures.py
47	dark windows with strong coupling	e47_dark.py, e47b_darkmap.py, explore/make_darkmap_figure.py
48	time-resolved formation across regimes	e49_time.py
49	anomaly ranking	e50_anomaly.py
50	prior art, overlap regime	—
51	summary of the overlap-regime project	e43_phase.py, e50_anomaly.py
52	effective 2×2 Hamiltonian	e52_heff.py
53	phase-average convergence	e53_phaseavg.py check, e53b_map_check.py, e53c_boundary.py, e53d_affected.py, explore/make_phaseavg_figure.py
54	exact phase-averaged map	e54_exactmap.py, e54b_width.py base eps map res, e54c_summary.py
55, 56	acoustic walls vs stacking-fault theory	e55_fault_check.py, e39_walls.py
57	Chalise et al., what "spectral weight" means	—
58	first Chalise reference case (invalid)	e58_chalise_ref.py
59	absorbing reference, closed-form transfer function	e58_chalise_ref.py, e59_transfer.py
60, 61, 62	literature status, quasi-static number	—
63	surface acoustic waves	e63_saw.py
64	Sondhauss & Wark (2003) vs Entry 59	e64_sw_compare.py
65	transfer function for Si 400	e65_si400.py [tt] [fig], e65_conv.py, e65_packet.py
66	Chalise Fig. 2 tails in full TT	e66_harmonic.py
67	overshoot map	e67_overshoot_map.py, e67_worker_audit.py, explore/make_poster_figs.py
68	one-number linearity limit (falsification run)	e68_nu_collapse.py run stats fit2 fig (PR #16)
69	S&W vs Entry 59 term by term	e64_sw_compare.py (PR #16)
70	signal max vs Fisher-information max	e70_fisher.py run fig (PR #16)
71	H needs a long window	e71_window_H.py (PR #16)
72	histories that cannot be told apart	e72_history_degeneracy.py (PR #16)
73	what S = 0.5-6 means physically	e73_realisability.py (PR #16)
74	directive questions B3, B4 (no computation)	—
75	H as a closed form in window length	e75_window_cf.py run posthoc fig (PR #16)
76	degeneracy class at larger S	e76_degeneracy_S.py runA runB fig (PR #16)
77	anomalous absorption χ_ih	e77_chi_ih.py run ttp ttm conv60 fig (PR #16)
78	±η side asymmetry	e78_side_asymmetry.py run fig (PR #16)
79	coupling ratio R and the degeneracy class	e79_degeneracy_R.py run fig (PR #16)
80	corrections to Entry 73	—
81	asymmetric Bragg as a rescaling	e81_asym_map.py run (PR #16)
82	beam divergence and H	e82_divergence.py run fig (PR #16)
83	beam condition at GHz-train frequencies	e83_lowx.py run fig (PR #16)
84	beam divergence and the centre formation law	e84_satellite_divergence.py run fig (PR #16)
85	Si 440 / 17.48 keV, 1-2 GHz case	e85_si440_case.py (PR #16)
86	integrated population and the beam condition	e86_population_divergence.py run fig (PR #16)
87	Laue integrated population at 3 w_n	e87_population_laue.py run fig (PR #16)
88	polarisation and Debye-Waller as rescalings	e88_polarisation_dw.py (PR #16)

Script names follow the entry numbers except in 42 (e46_bands.py), 48 (e49_time.py) and 49 (e50_anomaly.py). Not cited by any entry: explore/calculate_rk45.py (see §1), explore/make_collapse_figure.py, explore/make_dephasing_figure.py, explore/make_poster_qr.py, explore/p0_equations.ipynb (static penetration-depth table from the initial commit) and explore/nb/paper_figs.py (paper figures, notes/paper_draft.md §3).

3. Claims

Current table

This is notes/paper_draft.md §4 as of ffa4d9c (2026-10-23, research/paper-viability), copied verbatim; §4 is authoritative and this copy is not re-judged. "Confirmed" there means confirmed in the model by full TT or a closed form, and, for the rows the physics lane reviewed, cross-checked between AI lanes; it is not an experimental confirmation. No long-window satellite measurement exists.

#	claim	status	evidence
1	First-order linearisation of TT for a phonon	prior art	Sondhauss & Wark 2003 eq. 25–29 (Entries 64, 69)
2	Closed-form H with the Bloch-wave (Riccati) depth dependence, valid inside the Darwin width where the plane-wave form is off by 3–8×	confirmed in the model; novelty bounded by unread Chukhovskii & Förster 1995, Wark & Lee 1999	Entries 59, 64, 65
3	H needs a long window: the strain must travel ≥ max(1/κ, x/π) Λ during it; closed form H(f; T_w) exact to first order at every window 0.14–300 ns	confirmed (TT 0.3 %); packet-free short form rejected below ~100 τ	Entries 71, 75
4	Design numbers Si 400 / 17.48 keV: T_flat 19.6, T_half 32.8, T_90 52.2 ns; ripple with period 2x Λ/v	confirmed (TT to 0.002)	Entry 75
5	Chalise et al. 140 ps window sees no crystal factor (H flat to 4 %); their low-frequency deficit is only partly the second-order term K = 0.65–0.88	confirmed in the model; their normalisation unknown; any real beam also flattens the crystal factor (Entry 82)	Entries 66, 71, 82
6	χ_ih is a second tabulated crystal parameter: 0.9 % steady, 1.4 % at 52 ns, 3–4 % at 8 keV	confirmed; pre-registered rule fired	Entry 77
7	±η side asymmetry, strain-spectrum-free, reads χ_ih/χ_i0 (1.9–6.2 %, sensitivity 0.02–0.07)	predicted and confirmed in the model (TT 0.06 %); unobserved; a single-ray observable: odd in δ, it averages to zero over a symmetric beam (Entry 82)	Entries 77, 78, 82
8	Asymmetric Bragg = exact rescaling; one geometry factor g(√b + 1/√b)/2; T_90 shortens with Λ_asym	confirmed (TT 1e-16)	Entry 81
8b	Beam divergence / bandwidth: for σ_η ≫ κ + π/x the crystal factor is flat, κ√(π/2)/σ_η, at every window (x ≳ 10); H(f; T_w) and the ±η asymmetry are single-ray (meV, sub-µrad) observables; at x = 1–5 the tolerance is 0.3–0.6 µrad / 4–7 meV (8 keV) and the beam-averaged factor is non-monotonic in σ	confirmed (TT beam average vs closed form 0.02–0.24 %)	Entries 82, 83
9	ν ≲ 0.1 ⇒ weight linear to 2 % (1 − K ≤ 1.8 ν²)	confirmed as a bound on the scanned domain (Bragg, x ∈ {2, 5}, one packet family, weight at the peak)	Entry 68
10	"One dimensionless criterion joins the regimes"	falsified (spread ×10–14 at fixed ν)	Entry 68
11	Formation law I_n = sin²/tanh²(πα|J_n|t̃), Bessel ladder, period-ratio calibration	confirmed; prior art in kind DeCamp 2001 (n = 0), Polikarpov 1987/88 (static); as a centre observable it needs the beam condition of 12b	Entries 1, 9, 13, 19, 24; tests
12	Order n reads the one-transit running mean of J_n(S A); zero-mean perturbations degenerate (≤ 3e-6 for S ≤ 6); boundary R = 0.2	confirmed with a boundary	Entries 72, 76, 79
12b	Beam condition for the centre formation law: σ_η ≲ 0.3 w_n (0.1–0.3 µrad, Si 440 / 17.48 keV Laue; 0.6–1.4 µrad, Si 400 / 8 keV Bragg); at σ_η = w_n the first maximum loses 13–40 %	confirmed (full TT beam average, 8 cases)	Entry 84
8c	Polarisation: π = σ form with α → Cα (κ fixed), unpolarised = two-term closed-form sum (TT 0.02 %); negligible at Si 400 / 8 keV (C² = 0.12), 20–32 % at Si 440 / 17.48 keV; static Debye–Waller already in the X0h constants	confirmed	Entry 88
12c	Integrated Laue satellite population: flat-beam shape from σ_η ≈ 5 w_n (2–5 µrad, step) and 10 w_n (10 µrad, 1 t_tr ramp) with a one-channel detector (9–18 µrad); formation curve from TT (tabulated; 13–25 % off ∫J_0); Bragg population not clean	pre-registered threshold falsified: 3 w_n failed in 3 of 5 runs (Entry 87; Entry 86 failed under its own rule). 5 w_n (step) and 10 w_n (ramp) are post-hoc conditions from the same cases; confirmation needs a pre-registered run on new cases (T2s)	Entries 86, 87
13	Signal maximum = Fisher-information maximum for a ps packet at one angle	negative result	Entry 70
14	Bessel-zero resurrection at third order	confirmed in the model, unobservable (4e-5); unread Entin, Iolin–Entin	Entries 45, 51, 52, 54 (not in this paper; P4)
15	Realisability: S ≤ 6 at 1–2 GHz clean on the elastic side; drive power 6–900 W/cm² peak	constants only; harmonic column withdrawn (shear symmetry); transducer number unverified	Entries 73, 80

Withdrawn, falsified or corrected

Hypotheses in this list are not to be reopened without new evidence (tratt brief, BACKLOG T1).

Where	What	By
Entry 12	"the peak position shifts dynamically"	withdrawn: kinematic interference (Entry 3)
Entry 12	"the ±n asymmetry is a transit/dynamical effect"	withdrawn: a first-acoustic-wavelength effect, ∝ 1/α_ac (Entry 8)
Entry 12	"no additional timescale" (session 1)	withdrawn: t_tr/α_n exists
Entry 9	t_f ~ 1/α in Bragg	corrected to 1/α_n
Entry 31	grid rule	bug found and fixed in the entry
Entry 41	Entries 38-40 (chirp, walls, anomaly search)	"nothing qualitatively new": each effect has a known analogue
Entry 51	dynamical localisation or diffusion in order space; single-phase breakdown metrics	absent; micromotion that averages away
Entry 56	any "new mechanism" claim for the Entry 39 phase wall	withdrawn: it is a stacking fault per satellite (Authier & Simon 1968)
Entry 65 §2, Entry 66 §3	"17 keV under-reads 20 GHz by a factor 2" for the Chalise packet; "two factors multiply, 0.33-0.45"	withdrawn for the 140 ps window by Entry 71; holds only for long windows (the strain travels ≥ max(1/κ, x/π) Λ during the window; a 10 ps packet read over 52 ns shows it, Entry 75)
Entry 68	one dimensionless number sets the linearity limit (directive B3, strong form)	falsified: 1 − K spreads ×10-14 at equal ν
Entry 70	directive B2 (crystal moves the signal maximum away from the information maximum)	negative in the case computed
Entry 73	ε2/ε column and harmonic verdicts for the shear geometry	withdrawn by Entry 80
Entry 75	packet-free short form of H(f; T_w)	rejected below ~100 τ
Entry 65 §3(a)	the tt run at 24 RK4 steps per cycle (x ≥ 30)	not converged; "do not quote it" (48 steps per cycle is converged to < 0.5 %)
Entry 73 §1	"u0 = 7.6 pm × S"	corrected to 15.3 pm × S by Entry 80 (7.6 is the S = 0.5 row)
Entry 82	"an ordinary beam sees the kinematic law × a flat factor"	limited by Entry 83 to σ_η ≫ κ + π/x (x ≳ 10, the ps-packet band); at GHz-train frequencies the beam-averaged factor depends on x and σ
Entry 82	"1e-4 bandwidth = 35 µrad-equivalent" (angle and bandwidth added in quadrature)	corrected in the Entry 82–83 addendum (c): after a crystal monochromator angle and energy are DuMond-correlated and σ_η must be computed from the optics
Entry 85	first version with shear v_T	withdrawn, replaced by longitudinal [110]
Entries 86, 87	integrated population beam-tolerant at 3 w_n	Entry 86 failed under its own rule (amplitude metric error recorded); Entry 87's pre-registered 3 w_n failed in 3 of 5 runs

4. Reproducing the main computations

git clone https://github.com/NaNIsaac/TraTT.git
cd TraTT
uv sync --extra dev
uv run pytest -q         # solver tests: analytic limits, Floquet cross-check, formation law

Poster-era entries (0-67) run on main:

uv run python explore/nb/e1_order.py          # Entry 1, formation order
uv run python explore/nb/e9_bragg.py          # Entry 9, Bragg
uv run python explore/nb/e59_transfer.py      # Entry 59, closed-form transfer function
uv run python explore/nb/e65_si400.py tt fig  # Entry 65, Si 400 with the full-TT check
uv run python explore/make_poster_figs.py     # poster figures (poster/ is frozen)

Paper-viability entries (68-88) need the branch:

git switch research/paper-viability
uv run python explore/nb/e68_nu_collapse.py run stats fit2 fig
uv run python explore/nb/e75_window_cf.py run posthoc fig
uv run python explore/nb/e82_divergence.py run fig
uv run python explore/nb/paper_figs.py fig1     # likewise fig2, fig4b

Run times and convergence checks are in each entry. Several sweeps take minutes; on the shared machines they are scheduled as heavy runs. These commands were collected from the entries and docstrings; they were not rerun for this document.

5. Open questions and implementation tasks

From notes/paper_draft.md §5 and §7, notes/paper_candidates.md and the tratt handoff; listed, not re-ranked.

• Novelty of claim 2 depends on three unread papers (Sondhauss & Wark 2003 in full, Chukhovskii & Förster 1995, Wark & Lee 1999) and six more named in the cross-review (T2d, waiting for the PDFs).
• Paper scope: P1 + P2 in one paper or P1 alone; whether to propose the ±η asymmetry as a measurement; asymmetric cut; whether to accept the 5 w_n condition (Isaac's decisions, tratt handoff).
• Claim 12c needs a pre-registered confirmation run on new cases (T2s).
• Model limits stated for the paper: two-beam, plane wave, quasi-static, centrosymmetric; Laue lossless only; strain as a phase only (no elasto-optic change of χ_h); first order in S for the closed form.
• Experimental side: no long-window satellite measurement; transducer peak power unverified.
• Future project named in Entry 51: a perturbative theory of the overlap regime and its comparison with the Entin / Iolin-Entin resonance theory.
• Missing sources for the lineage itself: the FTT package (src/ftt/) and AW_TT_EXP/spec_calc.py were not found; if they exist on another machine, their place in §1 can be checked against code.