E.15:11 - SoTA-Echoing
The selected method combines exact edition comparison with dependency-aware incremental reverification and optional design-space search. No one source supplies the complete FPF procedure.
| Practice or research line | Status and contribution | Limit or rejected overread | Change to E.15 |
|---|---|---|---|
| Git diff documentation, current official documentation | Adopt: compare exact endpoints and keep historical values recoverable. | A textual or blob diff locates change but does not prove semantic preservation or actual consumer impact. | Exact predecessor/candidate recovery is mandatory; semantic probes remain separate. |
| Semantic Versioning 2.0.0 | Adapt as established release practice: immutable released contents and explicit compatibility communication are useful. | SemVer depends on a declared public API and its labels do not decide FPF semantic continuity. It is not the current method for impact analysis. | Version syntax is optional and follows, rather than supplies, the Delta-Class decision. |
| Current Bazel action-graph and incremental-build practice and its actual-versus-declared dependency distinction | Adapt: trace changed inputs through actual dependencies and reuse unaffected work. This advances the effort-to-reliability frontier over full reruns when change is local. | A build graph is not an ontology of pattern meaning, and declared references can miss or overstate semantic dependence. | Inspect actual and declared consumers; reopen only affected conclusions. |
| Li, Chen, Huang, and Ding, “Change-aware model checking for evolving concurrent programs based on Program Dependence Net”, 2024 | Adopt the current research move by analogy: use prior verified results and property-relevant dependency slices rather than rechecking an entire changed system. | Software paths and LTL properties do not transfer as FPF kinds or sufficient evidence for prose semantics. | Result reuse requires an unchanged conclusion and a justified outside-impact dependency. |
| MAP-Elites and quality-diversity search, with current FPF C.18/C.19 machinery | Retain as optional lineage and method family: useful when several materially different candidates and diversity itself matter. | Candidate multiplicity and archive coverage do not improve an understood local repair and do not select a winner across heterogeneous qualities. | Alternative generation is conditional; E.21/E.22 and intended use decide among non-dominated candidates. |
| Full rerun of every authoring, source, assurance, and review activity | Reject as the default rival: broad reruns can be appropriate after a genuinely broad Δ-3 change. | At comparable correctness, they spend more effort on unaffected premises and encourage ceremonial records. | Scope verification from actual change and dependencies, while preserving the explicit broad-change branch. |
The non-dominated contribution is therefore not a new authoring trace or scoring system. It is the combination of a cheap direct-repair path, actual-delta classification, independent predecessor preservation, actual-consumer reach, and whole-pattern practitioner-language verification, with stronger search and checking opened only by their real use.