Description delta reviewed

The check is description_delta_reviewed. It is VERIFY-stage. It errors. It fails the merge by default. It fires when four things are true together: the requirement is still FRETish-backed, the description changed on the branch, the FRETish text did not, and no no-formalization-change review was recorded. Informal shalls with empty FRETish are out of scope. It does not write the YAML. It does not prove the Go.

Changing the narrative without touching the formal model is sometimes correct. A typo. An example. Terminology. That judgement still has to be written down. Until it is, verification cannot treat the branch delta as reviewed. The two resolutions are alternatives, not steps. Record the review only if the behaviour is unchanged. If the wording now demands a new condition, a changed bound, a different response, or a widened scope, edit the FRETish instead and do not record the review.

The code-moved hop stays on authored delta expected. The stamp-over-new-export hop stays on no authored change surface reviewed. The backing hop stays on change evidence complete. A stale-but-unedited description that a parser outgrew stays on description grammar enumeration complete.

# SYS-REQ-210  fretish: when token.age > 30 then reject
#              description used to name the 30-second bound
# this branch rewrote the description: also drop revoked issuers
# FRETish still only names token.age
# a wording-only hop
# pass. the extra tests are green. the formula did not move
# proof audit --check description_delta_reviewed
# [VERIFY] description_delta_reviewed
# 1 changed FRETish-backed requirement descriptions lacked explicit formalization review
# SYS-REQ-210
# FAIL
# proof help description_delta_reviewed

Read the wording next to the formula before you reach for a stamp. Options 1 and 2 are alternatives, not steps. Do not record no-formalization-change when the intent actually moved. That review is a durable, signed claim that a human compared the new wording against the formal model and found them equivalent. Filing it over a real semantic change leaves the FRETish, the evidence, and every downstream result describing behaviour the spec no longer asks for, and the audit trail now says it was checked.

proof review req SYS-REQ-210
proof review impact SYS-REQ-210 --decision no-formalization-change --reason "Description clarified; FRETish and linked evidence remain valid"
proof req edit SYS-REQ-210 --fretish "<FRETish text matching the new intent>"
proof trace autolink
proof audit --check description_delta_reviewed --verbose
proof help description_delta_reviewed

The teaching graph is still one FRETish-backed requirement whose description moved and whose formula did not. Close it by recording no-formalization-change only if the behaviour is unchanged, or by editing the FRETish (and refreshing evidence links if the tests no longer witness it) if the intent moved. Do not record the review to make the finding disappear. The hop does not write that review for you. The hop does not edit the FRETish for you.

# SYS-REQ-210 fretish now names revoked issuers
# proof audit --check description_delta_reviewed
# 0 description findings
# VERIFY may move on. pass sits on a formula that moved with the wording

The code-moved hop stays on authored delta expected. The rewrite hop stays on characterization testing and mirrors. Jama still authors. Proof vs Jama.

Error when a FRETish-backed requirement's description changed on the branch, the FRETish and formalization strategy did not, and no no-formalization-change review was recorded. Default severity fails the merge. Informal shalls with empty FRETish are out of scope. A description that moved with the FRETish does not fire. Cannot-diff against the base branch is a pass, not a proved graph. Zero changed requirements is a pass. A recorded review is a signed claim, not a proof that the wording and the formula are equivalent. The hop does not write the impact review. It does not edit the FRETish. It does not write a description. It does not prove the Go. A quiet hop is not a proof that Jama's shall matches the formula, only that every in-scope description-only edit currently has an explicit review or that nothing asked. We have not scored this floor against a frozen Jama pack or a VectorCAST corpus. The loss is named, not scored.

The code-moved hop stays on authored delta expected. The rewrite hop stays on characterization testing and mirrors. The engagement stays on software correctness audit. Jama still authors.

What is description delta reviewed? Same question. Same URL.

Is this description grammar enumeration complete? No. That hop reports an unedited closed list the parser outgrew. This hop reports a description-only edit on a FRETish-backed shall. See description grammar enumeration complete.

Is this authored delta expected? No. That hop reports a missing no-authored-change stamp on a moved production file. This hop reports a description-only edit on a FRETish-backed shall. See authored delta expected.

Is this no authored change surface reviewed? No. That hop reports a present refactor stamp over new public surface. This hop reports unmatched wording. See no authored change surface reviewed.

Is this change evidence complete? No. That hop is whether the declared change type still has its backing. This hop is the description versus the formula. See change evidence complete.

Do extra tests clear a description-only edit? No. A green suite is not a review of the new wording.

Does recording no-formalization-change always clear it? Only if the behaviour is unchanged. If the intent moved, edit the FRETish and do not record the review.

Does a quiet hop prove the FRETish matches the wording? No. The hop observes fingerprints and a recorded review. It does not prove the Go.

Does this finding fail the merge? Yes by default. The check keeps error severity.

Are informal shalls in scope? No. Empty FRETish is out of scope. The hop only watches FRETish-backed descriptions.

Does a description that moved with the FRETish fire? No. The finding clears because the formula moved with the wording.

Is this characterization testing? No. That page is rewrite-with-a-ledger. This hop is a description-only edit. See characterization testing and mirrors.

Is Proof an LDRA alternative for the instrument? No. LDRA still owns the avionics toolchain. Proof vs LDRA.

Is Proof a Jama alternative for the shall? No. Jama still authors. Proof vs Jama.