QF Solver 0.2.5a0 Owner Review template¶
Review identity¶
| Field | Value |
|---|---|
| Candidate SHA | PENDING |
| Evidence digest | PENDING |
| Review date | PENDING |
| Owner | PENDING |
| Proposed decision | PENDING |
Allowed decisions: more_evidence_required, accepted_with_recommendations,
accepted_for_bounded_experimental_use, accepted_for_release_0_2_5.
For the original unchanged scope, only accepted_for_release_0_2_5, after G11
passes on the exact candidate SHA, can close G12. If the Owner explicitly
approves a narrowed release scope, the decision must instead record
OWNER_SCOPE_REVISION = APPROVED and SCOPE_CHANGE = YES; G12 then applies
only to the remaining qualified scope and never promotes excluded gates.
The final 0.2.5a0 decision is recorded separately in
0_2_5_owner_scope_revision.md.
Planning approval questions¶
- Approve the MUST/SHOULD/COULD boundary and optional friction policy?
- Approve TET4/HEX8 as MUST and TET10/HEX20 as SHOULD for large deformation?
- Approve Total Lagrangian StVK as the elastic geometric baseline, subject to audit, or request a different formulation?
- Select the bounded J2+geometric model: corotational small-strain J2, a newly planned finite-strain plasticity model, or removal of J2+geometry from MUST?
- Approve one Crisfield spherical arc-length method as the initial continuation target?
- For frictionless contact, preserve the existing active-set multiplier method as reference, adopt penalty, or plan augmented Lagrangian from the start?
- Confirm Full Newton as the only production-qualified nonlinear strategy?
- Confirm the 0.2.4 coverage policy as the baseline rather than creating a new percentage target?
- Confirm Code_Aster as primary external reference and rules for N/A tools?
- Confirm that external numerical correlation is not a physical-validation claim?
- Confirm triple coupling and Coulomb friction do not block 0.2.5a0?
- Approve the gate dependencies and STOP/GO policy?
Final release questions¶
- Are all mandatory gate records complete on the candidate SHA?
- Are objectivity, consistent tangent and transactional-state proofs adequate?
- Are buckling and arc-length claims bounded to demonstrated problem classes?
- Is contact enforcement sensitivity acceptable and documented?
- Are coupled formulations mathematically approved and externally correlated?
- Are all 0.2.4 capabilities free of unexplained regression?
- Are performance and memory limits honestly stated?
- Do public docs avoid claiming unqualified friction/high-order/triple coupling?
- Do wheel/sdist/docs/smoke artifacts match the candidate SHA?
- Does the Owner authorize an Owner-controlled tag, GitHub Release and PyPI publication as separate operations?
Decision¶
Owner decision: PENDING
Conditions/recommendations: PENDING
Signature: PENDING