What is software change impact analysis, and how do I keep the blast radius true in CI?

Bohner's CIA is the process of finding what else a change touches. LDRA sells that as a capability on C and Ada. Jama stores a change-impact field on the requirement object. Wikipedia treats IA as consequences inside a deployed product. This page is the graph in the repo: which files, tests, and docs currently link to the shall, printed so a reviewer can read them, then a merge gate that refuses a traced edit with no recorded decision.

The matrix H1 lives on requirements traceability matrix. The LDRA H1 lives on Proof vs LDRA. The unprefixed query change impact analysis is mostly Prosci and Six Sigma. That is not this URL.

proof trace impact SYS-REQ-096 --format table
proof workflow check --stage implement --only authored_delta_expected --verbose
proof review impact SYS-REQ-096 --decision no-authored-change --reason "Warm-cache persistence changed implementation strategy only; requirement and linked design intent remain valid"

On this install the teaching pair is SYS-REQ-096 and a traced production file. The miss is the edit that never asked for a review, so CI could not fail when the shall and the file drifted.

The teaching pair is SYS-REQ-096 and a traced production file. The check looks at implementation files changed in the current branch. If the file is owned through implemented_by and neither the requirement nor a linked design artifact changed, the check warns until a no-authored-change review targets that file with matching fingerprints.

proof review req SYS-REQ-096
proof workflow check --stage implement --only authored_delta_expected --verbose
proof review impact SYS-REQ-096 --decision no-authored-change --reason "Warm-cache persistence changed implementation strategy only; requirement and linked design intent remain valid"

Do not record no-authored-change over a new flag, subcommand, grammar token, or any new externally-observable behavior. Those are authored-contract changes: update the spec, or file a problem report plus the regression test. Recording the stamp over a contract change is a false attestation. We have not run LDRA Testbed and Proof on the same frozen corpus, and we have not claimed the table is a static analysis of the binary. The loss is named, not scored.

The blast-radius command only sees links already in the graph. A file with no implemented_by edge will not appear. Overlay-audit projects do not use authored_delta_expected as a source-native no-authored-change gate. Tests alone do not clear the check for production-code changes. Description-only edits belong to description_delta_reviewed, not this stamp. Object-code coverage on C still wants the qualified toolchain. That H1 lives on MC/DC for Go and on Proof vs LDRA.

The matrix H1 stays on requirements traceability matrix. Jama still authors.

What is change impact analysis without the software prefix? Prosci, Six Sigma, and organisational change. Not this URL.

What is an impact review? The stamp, not the table. proof review impact records that the shall still stands. Affiliate-marketing SERPs for the bare phrase are not this cluster.

What is a requirements traceability matrix? The table of links, re-derived from the graph. RTM.

Does Proof replace LDRA CIA? No. LDRA still wins at C, C++, and Ada in a qualified toolchain. Proof vs LDRA.

Is Proof a Jama alternative for impact? No. Jama still authors. Proof fails the merge when a traced file moved and nobody recorded the decision.