Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions changelog.d/kringel-counters-diagnosis.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
### Added

- **A diagnosis of why the counters close at the plate nib, measured along
the whole chain.** `docs/notes/kringel-binnenflaechen-2026-09-06.md` takes
every counter of the nine loop-losing glyphs that the plate's ink and the
composition share, and follows it from the ink through the frozen skeleton
and the stored Kette v5 fit to the composed row — all as one comparable
quantity, the centerline loop aperture, from which a Gleichzug counter is
exactly `D0 - 2h`. (For `p` that shared counter is the large loop; its
collapsed shaft loop is a separate chart-versus-Laufform finding, read off
the stored rows rather than matched to the plate.) The composed loop is 0.10–0.20 xh
narrower in diameter than the plate's, and about three quarters of the
Kette's own share of that (per matched loop, median 0.75) is inherited from
the medial axis rather than made by the fit: where two pen passes merge
around a tight counter, `skeletonize` returns the axis of the merged blob
instead of the two paths. The stored row is narrower again than the
occurrences it stands for — in direction nine times out of nine, though the
note says plainly which comparison would be needed to charge that amount to
the estimator. The composer itself loses nothing (<= 0.0008 xh against the
row), so the two places worth an arm are the evidence and the estimator, not
the fit. Report only: nothing built, adopted or written, and the four
resulting candidates are named with a one-line pre-registration sketch each
in `tintenfolger.md` §7.9.
1 change: 1 addition & 0 deletions docs/index.md
Original file line number Diff line number Diff line change
Expand Up @@ -133,6 +133,7 @@ Nachschlagequellen, keine Pläne.
| [notes/serve-image-importgraph-2026-09-04.md](notes/serve-image-importgraph-2026-09-04.md) | Nachmessung zum Serve-only-Image: was `import api.main` lädt, was die Pakete wiegen, zwei Wege mit ihrem Preis — geändert wurde nichts (die eine offene Messung löst die Runde vom 2026-09-05 ein) |
| [notes/serve-image-importtime-2026-09-05.md](notes/serve-image-importtime-2026-09-05.md) | Zweite Runde dazu: dieselben Sätze in einer image-förmigen venv (ohne Extras, Bytecode vorkompiliert) plus die Aufteilung des Imports nach Modulen — das Urteil vom 2026-09-04 hält |
| [notes/eigenhand-router-importzeit-2026-09-05.md](notes/eigenhand-router-importzeit-2026-09-05.md) | Warum `api.routers.eigenhand` als teuerstes Modul galt: zwei Drittel der 64,9 ms sind eine GC-Pause, die `-X importtime` dem laufenden Modul zuschreibt — nichts zu verschieben |
| [notes/kringel-binnenflaechen-2026-09-06.md](notes/kringel-binnenflaechen-2026-09-06.md) | Warum die Kringel bei der Platten-Feder zulaufen: die Mittellinien-Schleife ist 0,10–0,20 xh enger als die der Platte — zerlegt in Skelett · Kette · Zeile, drei Hypothesen geprüft (zwei entschieden, H2 nur der Richtung nach), vier Rettungswege benannt (nichts gebaut) |
| [notes/quellen-recherche-2026-07.md](notes/quellen-recherche-2026-07.md) | Geschriebene Wortvorlagen und echte Hände: Rangliste, Absteiger, mögliche Anfragen |
| [notes/stifte-fuer-unterwegs.md](notes/stifte-fuer-unterwegs.md) | Stift- und Hardware-Recherche fürs Schreiben unterwegs |

Expand Down
Loading
Loading