Skip to content

0.2.6 G10-LOT1 - Adversarial Owner Review

Review state: PARTIAL; formal Owner closeout remains pending.

Execution source SHA: 51b3a7c8ace6731830109984a01ce31f79c44401. The review evidence is documentary and may be committed at a later SHA. It does not replace the execution source SHA or alter any numerical result.

Review conclusion

The ten Lot 1 route records were reviewed against their implementation path, existing tests, maturity, ownership and evidence boundary. The review found no real bug, no numerical regression and no reason to change G07, G08, G09, G11 or G12. Internal tests alone do not promote a research route.

The existing bounded small-strain J2 classification remains owned by G06. Transaction/checkpoint/retry and structured failure diagnostics retain their bounded G09/G11 ownership. Total-Lagrangian, arc-length, finite-kinematic J2 and coupled routes remain outside a new G10 qualified claim.

Route decisions

Route Review classification Reason and boundary
Small-strain nonlinear static OWNER_KEEP_EXISTING Retain the existing bounded G06 scope; no G10 expansion.
Total-Lagrangian elasticity OWNER_NEEDS_EXTERNAL_EVIDENCE G07 remains parked; compatible external evidence would directly address its existing blocker.
Arc-length continuation OWNER_NEEDS_EXTERNAL_EVIDENCE Internal branch evidence is insufficient; comparable continuation evidence has high value.
Finite-kinematic J2 OWNER_NOT_QUALIFIED Available external J2 decks are small-strain and are not an equivalent finite-kinematic oracle.
J2 plus geometry OWNER_BLOCKED No external deck currently separates constitutive and kinematic differences.
Geometry plus frictionless contact OWNER_EXPERIMENTAL_ONLY Existing contact comparisons use a different enforcement or surface model.
J2 plus geometry plus contact OWNER_BLOCKED It inherits unresolved finite-kinematic J2 and contact-equivalence blockers.
Modified Newton on finite routes OWNER_NOT_QUALIFIED The existing scope validator rejects this path for qualified finite-kinematic use.
Transaction/checkpoint/retry OWNER_KEEP_EXISTING Retain the bounded G09/G11 evidence and its stated coverage limit.
Structured failure diagnostics OWNER_KEEP_EXISTING Retain the bounded G11 evidence; exhaustive route coverage is not claimed.

External evidence selected

Only two routes are selected for a future targeted campaign. They are routed to G07 ownership and selection does not reopen or close G07:

  1. Arc-length continuation: use the existing two-TET4 shallow-arch/ snap-through case and a Code_Aster LONG_ARC deck with the same geometry, load convention, control variable and continuation direction. Compare the complete load-factor/displacement and reaction curves, turning point, residual history and restart branch. Existing G07 policies must be used; no new tolerance is approved by this review.
  2. Total-Lagrangian elasticity: use a same-mesh TET4/HEX8 structural case with equivalent Green-Lagrange/PK2 measures and a positive det(F) domain. Use Code_Aster or CalculiX only when the kinematics, material, loads, constraints and output measure are genuinely comparable. Compare complete load-displacement and reaction curves, stress/strain measure, energy, det(F) and residual. Non-comparable decks remain skipped.

These are the only selected routes because their existing external paths can test the unresolved question without changing QF Solver. The selection is a campaign recommendation, not a PASS and not an Owner approval of a new threshold.

Routes deliberately not selected

  • Finite-kinematic J2: the available Code_Aster structural J2 decks use DEFORMATION="PETIT"; the available CalculiX J2 check is a small-strain material-point comparison. A new external run would be misleading until a compatible finite-kinematic constitutive reference is identified.
  • J2 plus geometry: combining the existing small-strain J2 deck with the TL elastic deck would hide the formulation difference rather than measure it.
  • Geometry plus contact: the existing comparisons do not share an identical penalty/contact surface formulation, so a scalar reaction comparison would not isolate the coupled route.
  • Triple coupling: it inherits both unresolved blockers and is not a useful first external experiment.

Acceptance boundary

The selected campaigns must declare geometry, units, kinematics, material, loads, constraints, continuation controls, observables, tolerances and output digests before execution. Any missing or non-comparable condition is SKIP or OWNER_REVIEW_REQUIRED, never an inferred PASS. No universal tolerance is introduced here.

G10 disposition

026-G10 remains IN_PROGRESS. Lot 1 can be closed as an audit checkpoint, but the gate cannot be closed without a formal Owner decision and the evidence requested for any route that is to receive a stronger maturity classification. The next action is the two selected, G07-routed external campaigns, subject to Owner acceptance. Full regression remains skipped by the Lot 1 policy.