V0.26 Scope - KS Revisit Decision, Promotion Deferred to V0.26b#
Status: COMPLETE
Summary#
v0.26 is a Kuramoto-Sivashinsky revisit decision release.
It revisits the v0.11/v0.12 KS no-go with modern confidence diagnostics, but does not promote public KS runtime APIs.
Stable investigation path:
internal normalized KS fixture
-> spectral_fd derivatives through u_xxxx
-> internal KS residual feasibility
-> translation fit / verification / confidence report
-> explicit v0.26 decision label
Release conclusion:
current_no_go_reference_fallback
KS remains deferred because the primary frozen fixture still relies on reference-fallback translation fitting. Public KS promotion, if justified later, is reserved for a separate v0.26b scope freeze.
Decision Labels#
current_no_go_reference_fallback: residual and verification evidence are feasible, but direct residual-based SVD fitting still drifts and requires reference fallback.residual_feasible_fit_not_promotable: residual and verification evidence pass, but fit evidence is not promotable.direct_strong_candidate_for_v0_26b_promotion: the frozen primary fixture passes direct-SVD/no-fallback criteria and may justify openingv0.26b.deferred_no_go: core evidence is not sufficient for either residual feasibility or future promotion candidacy.
Frozen Decision Criteria#
Promotion candidacy in v0.26 requires all of:
residual max
< 5e-2residual RMS
< 1e-2first verification error
< 5e-4verification classification not
failedselected/SVD span distance within frozen tolerance
fit evidence label
direct_svd_in_tolerancereference_fallback_used is False
confidence_label == "strong" is necessary but not sufficient for future promotion. Direct SVD in tolerance and reference_fallback_used is False are required.
Minimal KS Reproduction Matrix#
v0.26 records test-only evidence for:
frozen primary fixture
one seed sweep
one fit-epsilon sweep
one resolution variant
fit/fallback diagnostics
verification diagnostics
generator confidence report
All matrix variants are diagnostic-only. Promotion evidence must come from the frozen primary fixture, not from the best-performing variant in a sweep.
V0.26b Follow-Up Contract#
v0.26b is reserved for the actual KS promotion release if a future scope freeze accepts the direct_strong_candidate_for_v0_26b_promotion evidence.
Candidate v0.26b surface, if opened later:
KS data generator under
pdelie.dataKS residual evaluator under
pdelie.residualsKS vertical-slice example under
pdelie.examples
No weak KS, custom KS coefficients, custom KS initial conditions, broad KS regimes, or root exports are included unless explicitly frozen in v0.26b.
Explicit Non-goals#
no public KS data generator
no public KS residual evaluator
no public KS vertical-slice example
no public KS status example
no residual-only KS API
no weak KS API
no custom KS initial-condition API
no configurable KS coefficient API
no broad KS regime support
no root KS export
no broad adapters or file loaders
no multidimensional or nonuniform stable support
no time-translation APIs
no neural or callable generator APIs
no operator-facing APIs
Milestone Status#
Milestone 0: COMPLETE
Milestone 1: COMPLETE
Milestone 2: COMPLETE
Milestone 3: COMPLETE
Milestone 4: COMPLETE
Milestone 5: COMPLETE
Milestone 6: COMPLETE
Release Gate#
v0.26 is complete only if:
the primary KS fixture records
current_no_go_reference_fallbackresidual and verification evidence remain feasible but fit evidence remains fallback-backed
the confidence report is included and does not override the direct-SVD/no-fallback rule
minimal matrix evidence remains diagnostic-only
no best-of-sweep promotion path exists
no public KS runtime API, weak KS API, status example, or root KS export lands
API_STABILITY.mdrecords the decision-only status without adding stable KS runtime contractsCI uses one compact current release gate plus full editable tests and package smoke
package-index publishing remains deferred until
v1.0or later