Dane Sabo 645f2d8d27 prompt-jump model + app v2 + overnight journal entry (in progress)
Singular-perturbation reduction of the PKE+T/H system: set dn/dt=0,
solve algebraically n = Λ·Σλ_i·C_i / (β-ρ). State drops 10 -> 9 (no
n), removes Λ⁻¹ stiffness. Validated against full state on the heatup
scenario:

  t [s]    |Δn|/n_full   T_c err [K]
  60       3.7e-5        4e-6
  300      3.8e-4        1.9e-4
  1200     1.0e-3        2.2e-3
  3000     5.0e-4        7.2e-3

Maximum relative error 0.1% on n, peak 7 mK on temperatures over
50 minutes.  PJ approximation is excellent for slow heatup transients
(sub-prompt-critical regime).

Files:
  - code/src/pke_th_rhs_pj.jl: reduced 9-state RHS
  - code/scripts/validate_pj.jl: side-by-side sim
  - code/scripts/reach_heatup_pj.jl: TMJets reach with PJ model
    (probing T = 60, 300, 1800, 5400 s)

App v2 (Pluto):
  - §9b: live ingestion of reach_operation_result.mat with per-
    halfspace margins computed from JSON-defined inv2_holds.
  - §9c: 2D projection chooser (n, T_f, T_c, T_cold) with reach
    tube envelope overlay.
  - §9d: PJ heatup reach summary (placeholder until first run lands).

Journal:
  - Added 2026-04-20-overnight-prompt-jump.tex with PJ derivation,
    validation table, soundness ledger update.  apass markers for
    the in-progress reach results.

This commit captures state mid-run; next commit will add the
populated reach results once TMJets returns.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-20 22:45:24 -04:00
..

Lab Journal

                        ___
                      ,'   `.
                     /  ___  \      n, C_i,
                    | ( UO2 ) |     T_f, T_c, T_cold
                     \  ‾‾‾  /         |
                      `.___,'          v
                       |||||  →→→→  [ controller ] →→→→ rods
                       |||||
                  ‾‾‾‾‾‾‾‾‾‾‾‾‾  primary loop
                  ___________
                 |   SG       |     Q_sg(t) ∈ [Q_min, Q_max]
                 |____________|

The HAHACS invention log for the pwr-hybrid-3-demo preliminary example. Each dated session gets an entry. The goal: a reader in 2030 should be able to rebuild the thesis work from this journal alone.

Structure

journal/
  preamble.tex         shared LaTeX setup (fonts, listings, callouts, macros)
  journal.tex          top-level aggregator (builds all entries into one PDF)
  entries/             dated session entries, one file per session
    YYYY-MM-DD-{topic-slug}.tex
  figures/             journal-specific figures (referenced from entries)
  README.md            this file

Conventions

Filename

entries/YYYY-MM-DD-{topic-slug}.tex. One file per session. If a day has multiple distinct sessions, use time-of-day in the slug (e.g., 2026-04-20-morning-predicates.tex and 2026-04-20-evening-mega-session.tex).

Entry skeleton

\session{2026-04-17}{duration}{one-line summary}

\section{Session: ... (YYYY-MM-DD)}
\label{sec:YYYYMMDD}

\subsection*{Goal}
What I set out to do and why.

\subsection*{What landed}
...

\subsection*{Key decisions}
...

\subsection*{Dead ends}
...

\subsection*{Derivations}
...

\subsection*{Results}
...

\subsection*{Limitations recorded}
...

\subsection*{Open at close}
...

Entries compile standalone (each starts with \input{../preamble.tex} wrapped in a conditional so it only pulls preamble when not already loaded by journal.tex), or together via journal.tex.

Two entry styles

  • A-style (deep / invention-log): full derivations, code commentary, dead-ends, embedded figures, terminal output where useful. Used for retroactive entries and sessions that land meaningful artifacts.
  • B-style (narrative + pointers): end-of-session notes. Uses \apass{...} callouts to flag spots that need a later A-pass.

Callout boxes

From preamble.tex:

Environment Use
derivation Math derivations
decision Design choices with rationale + alternatives
deadend Paths that didn't work
limitation Soundness gaps, known-approximate behavior

Plus the inline \apass{text} marker for A-pass TODOs.

Code inclusion

  • \juliafile[options]{path/to/file.jl} — includes a Julia source file as a numbered listing.
  • \matlabfile[...]{...} — for MATLAB sources.
  • Or inline with lstlisting environment and language=Julia for snippets.

Always include the path as the listing caption so readers can find the file.

Figures

Figures live in ../docs/figures/ (shared with the thesis) or figures/ (journal-only). The preamble sets \graphicspath to check both.

Always include:

  1. A descriptive caption (what's on axes, what's being shown).
  2. A discussion in the surrounding prose — what the figure proves or illustrates. Figures without discussion are noise.

Terminal output

For a numerical result or an error that drove a decision, include the actual terminal text in a lstlisting block with style=terminal:

\begin{lstlisting}[style=terminal]
TMJets: 10583 reach-sets
T_c envelope: [274.45, 295.0] C
FAILED: AssertionError: radius must be nonnegative
\end{lstlisting}

Don't include full logs — only the lines that changed what you did next.

Build

cd journal
latexmk -pdf journal.tex            # whole journal as one PDF
latexmk -pdf entries/2026-04-17-controllers-linear-reach.tex  # one entry

Requires TeX Live with tcolorbox, listings, inconsolata, siunitx, cleveref, hyperref, fancyhdr. All in the standard distribution.

Not a replacement for

  • claude_memory/ — short AI-context notes, Markdown, different audience.
  • reachability/WALKTHROUGH.md — standalone doc summarizing current state of reach analysis.
  • Git commit messages — per-commit rationale for code changes.

The journal is the chronological narrative of the work. The others are snapshots, summaries, or pointers. They're all legitimate; they do different things.