Skip to content

Log

Entries are the dated events the notes record, newest first; git holds every earlier edit.

2026-11-01

2026-10-30

  • Update: The no-MPI claim is narrowed to the lock file and the build configuration, and the rectification sentence no longer asserts that a fixed-charge asymmetry alone rectifies (WP57; REV-111, REV-155). (1. FEM backend, 7. Ionic current rectification)
  • Update: The notes on the import cost and on the stage workspace name the Python interface, Case and Session, in place of load_case, resolve and run_case (WP57). (5. GUI and packaging, A stage's workspace)

2026-10-29

  • Update: Rectification needs an axial asymmetry of the geometry or of the fixed charge, not of the geometry alone (WP55's physics review of the sweep session). (7. Ionic current rectification)

2026-10-24

  • Update: Added the offscreen Qt and headless Playwright section to 07-software-stack/05 and marked the libEGL shim as superseded (WP5.01). (5. GUI and packaging)

2026-10-21

2026-10-19

  • Update: Marked the advice to mollify the wall-distance field and the advice to apply field-wise row scaling as superseded by the CODE_REVIEW_010 rulings (PHY-02, NUM-31, NUM-10). (1. Symbols, 7. Implementation notes carried forward, 3. Nondimensionalisation)
  • Update: Stated that the validated wall-distance field is the unsmoothed C⁰ field and that C¹ mollification is a deviation, off by default (author ruling, CODE_REVIEW_010). (8. Mesh resolution and adaptivity)
  • Update: Retargeted the [tested: …] markers that named a test deleted by the test scope (WP50, §8.2.13 N3): 31 markers became [verified] on their source footnote, or were dropped where the concept has none, and pending.yaml was removed.
  • Update: Removed the empty **** markers that retargeting left in nine concepts, the empty verified: key of concept 13, and the present-tense statements about --extended, no_cover and test_plan_shape.py that WP50 made false (WP50 review).

2026-10-18

  • Update: Stated that ngsolve/__init__.pyi ships only in the Linux wheel, so the stub test checks it where present. (16. NGSolve's bundled stubs cannot type)
  • Update: Repointed the six source citations that named plan files which do not exist (wp18, wp20, wp21, wp28, wp30, wp32) at the plans that do, and re-wired the claims for the 1 M NaCl conductivity, the zero-bias route check, the telescoping round-off statistics, the trailing-NUL behaviour and the fluid-only indicator to tests that assert them (WP48 review).
  • Update: Corrected the integration-order example to the tested fourth-degree polynomial and synchronised the specification's maintenance text with the closed, zero-pending ledger (WP48 review).
  • Update: Linked executable force-scale arithmetic to its existing assertions, recorded the exercised gemmi version, and restricted the FEM wheel table to its checked lockfile range (WP48 review).
  • Update: Repaired historical plan addresses and viewer attribution, pinned hashing, singular-system and sweep measurements to their immutable records, and separated historical magnitudes from current reproductions (WP48 review).

2026-10-16

  • Update: Backed all 63 remaining domain claims across the bundle with Tier-1 and Tier-2 test evidence and immutable plan records, reducing total pending claims across the knowledge bundle to zero and lowering the ceiling to 0 (WP48 item 8).
  • Update: Backed all 51 pending NGSolve-trap Library Behaviour claims in 06-numerics-fem/09-ngsolve-traps-found-by-implementing.md with Tier-1 and Tier-2 test evidence (tests/tier1/test_ngsolve_traps.py and existing suites) and historical commit footnotes, lowering the pending ceiling from 114 to 63 (WP48 item 7).
  • Update: Backed all 90 pending software-stack Library Behaviour claims across 07-software-stack/01..16 with Tier-1 and Tier-2 test evidence (tests/tier1/test_library_behaviour.py and existing suites), lowering the pending ceiling from 204 to 114 (WP48 item 6).
  • Update: Recorded tested_with package-version mappings across all seventeen Library Behaviour concepts and enforced non-empty tested_with mappings in VER-68 (WP48 item 5).
  • Update: Replaced every executable [verified] arithmetic and closed-form claim across the bundle with Tier-1 test evidence (tests/tier1/test_knowledge_arithmetic.py and existing Tier-1 suites), reducing verified_unsourced to zero (WP48 item 4).
  • Update: Backed all 100 pending Test Evidence claims across the seventeen Test Evidence concepts with explicit test assertions or immutable historical records (WP48 item 3).
  • Update: Linked source footnotes and immutable plan records to report readings, topological/calculus identities and standalone archive observations across the bundle, including the ten unsourced COMSOL reference-settings claims (WP48 item 2).

2026-10-14

  • Update: The folder lost its leading dot and is now knowledge/; every citation follows.
  • Update: The notes became a bundle of concepts, one idea each, with a header, a ledger of test-backed claims and this log (WP47).
  • Update: Recorded how the documentation build treats the bundle (13. Documentation tooling: MkDocs, Material and Zensical).

2026-10-12

2026-10-06

  • Record: After import nanopnp.io.case the only nanopnp subpackages in sys.modules are core and io , 6 October 2026 (5. GUI and packaging)

2026-10-05

2026-10-04

2026-10-03

2026-10-02

2026-10-01

2026-09-30

2026-09-29

  • Record: The geometry pipeline as executed, with measurements on the MD archive, the contour script and 2WCD. (1. Geometry pipeline, as executed)
  • Record: DensityMap.read exported .map and failed to read it back until WP25 did so , 29 September 2026 (2. Structure, trajectory and density)
  • Record: Gmsh facts, measured on 4.15.2 with netgen 6.2.2606, Linux (WSL2), 29 September 2026 (WP23; the implementation is mesh/gmsh_backend.py, the evidence the WP23 plan's Design and Outcomes) (4. Geometry, meshing and I/O)
  • Record: Measured 29 September 2026 on a Linux rebuild of the probe recipe in this container (WP24 D16, D17): PyInstaller 6.22.3, MDAnalysis 2.10.0, gemmi 0.7.5, scikit-image 0.26.0, Shapely 2.1.2 with GEOS 3. ... (5. GUI and packaging)

2026-09-28

2026-09-26

2026-09-25

  • Record: The author ruled on 25 September 2026 that R_i is the CHARMM Rmin/2 of PDB2PQR's CHARMM.DAT, the set carried by the reference ensemble's per-frame PQR files (§1.1) (1. Geometry pipeline, as executed)
  • Record: The structure, trajectory and density libraries and how they behave. (2. Structure, trajectory and density)
  • Record: That copyleft is file-level, so it is fine used unmodified as a separate package (CON-09, amended 25 September 2026) (6. Licence summary)

2026-09-24

2026-09-23

2026-09-22

  • Record: The GUI and packaging stack: PySide6, the web viewer and PyInstaller. (5. GUI and packaging)

2026-09-21

  • Record: Measured 21 September 2026 on the default ladder of an epnp-ns case at 0.1 M and 20 mV, twelve rungs on a cheap cylindrical pore (5. Nonlinear solution)
  • Record: Measured 21 September 2026, P2 GridFunction on a unit-square mesh, Draw(gf, mesh, order=k, show=False).GetData() then json.dumps (5. GUI and packaging)

2026-09-20

  • Record: Chosen for the bundle, one-dir (ADR-004, amended 20 September 2026) (5. GUI and packaging)

2026-09-06

  • Record: Measured directly (6 September 2026; too large to vendor, so it is named by $NANOPNP_REFERENCE_DATA and Tier 3 skips without it) (3. Fixed-charge pipeline)

2026-09-02