Review protocol and evidence status

Page status: generated scoping-review snapshot and evidence-status record.

This page is generated from review/snapshot-manifest.json and the canonical evidence matrix. It publishes the review state without implying that a single-coded seed corpus is an independently validated systematic review.

Published snapshot

fieldvalue
protocol version0.1.0
snapshot date2026-08-25
matrix records37
matrix SHA-2563cbb795c0cd6063da9da5979056c04c294ebe6ae8bc7545f1b0a708dc39ee9f8
deduplication rows37
deduplication SHA-25675143b8c286bbba4bd674c11fb205d594a1b7385783d634de0d11dcf79477b86
independent human double-codingno

Screening and coding counts

dimensioncounts
screening statusinclude: 36; exclude: 1; uncertain: 0
coding statussingle_coded: 37
exactness labelsexact: 14; not_reported: 4; outer: 1; scenario_approximate: 8; unclassified: 10

The current snapshot is therefore a single-coded seed snapshot, not a double-coded corpus. The 2026-08-15 second-coding log recommends eight rows for promotion after slot repairs and identifies six substantive coding conflicts; those recommendations remain pending and are not silently reflected in the canonical matrix.

Search coverage

The protocol names the following information sources:

  • IEEE Xplore
  • Scopus
  • Web of Science
  • Engineering Village/Compendex
  • Inspec
  • MathSciNet or zbMATH where available
  • arXiv
  • relevant standards and official software documentation

The dated search runs included in this snapshot are:

The search-run files record query families, available platforms, and limitations for each run. The protocol also requires backward and forward citation chasing, duplicate resolution, and a saved-search rerun before a tagged release; these are not claimed complete merely because a row is in the matrix.

Evidence-status interpretation

A populated matrix row means that one source-to-target result was coded at a declared scope. It does not mean that the source is a proof of the book's architecture, that an exactness label applies to a feasible set, or that a software conversion preserves provenance and limits. The exactness_object, recovery_map, constraint_map, and provenance_map fields must be read together.

The second-coding and independent-technical-review documents are retained as review evidence, but they are not substitutes for an independent human double-coding pass. The canonical matrix remains the source of published counts; dated snapshots and reconciliation notes explain proposed changes without rewriting history.

Reproducibility inputs

The manifest records 9 hashed protocol, coding, bibliography, and search-run inputs. Run scripts/check_review_snapshot.py to verify the matrix, deduplication register, record identifiers, and input hashes against this published snapshot.

The protocol's promised PRISMA-style flow is represented here as a scoped count table rather than a clinical-review claim: the available evidence is a seed search with one explicit exclusion, and the screening/coding pipeline is still being expanded.