Skip to content

Target-Space Difference Review

Change-triggered coverage review — instantiates Target-Complete Mapping Design

Diffs the required target set and mapping against their last certified baseline and treats every newly added or changed target as uncovered until a fresh witness proves otherwise.

Coverage is a property that rots. Target-Space Difference Review compares the target space and mapping against their last certified baseline, isolates what was added, split, redefined, or whose source changed, and reopens coverage for exactly those deltas. Its defining move is that a newly appearing target does not inherit green status: the diff makes each added or changed target uncovered-by-default until it is re-witnessed. That single default is what keeps a once-complete map from silently going false as obligations, definitions, and sources drift underneath it.

Example

A bank's regulatory report was certified complete last quarter — every mandatory field had a validated data derivation. Mid-quarter the regulator adds three new disclosure fields and redefines a fourth. The difference review diffs the new obligation set against the certified baseline: the three added fields have no derivation yet, and the redefined field's existing witness now targets a stale definition. All four flip to uncovered. The review also catches a source-side change the obligations diff alone would miss — a decommissioned upstream table has silently orphaned two previously-covered fields. The outcome is six reopened targets routed to owners before the next filing claims a completeness it no longer has.

How it works

What distinguishes this review is the baseline diff with an uncovered-by-default rule, not a full re-audit. It snapshots the certified target set and mapping; on each trigger — a new obligation, a target split, a source retirement, a mapping edit, an evidence expiry — it computes the delta against the baseline, marks added and changed targets (and newly orphaned ones) as uncovered, and requires impact review before any source or relation change is approved. The cadence at which the diff runs — periodic, event-driven, or risk-based — is set to match how fast the target space actually moves.

Tuning parameters

  • Trigger set — which events fire a review: target changes only, or source, dependency, and evidence-expiry events too. Narrow triggers miss silent drift.
  • Cadence — fixed schedule vs event-driven vs risk-based. Faster catches drift sooner at higher effort.
  • Baseline granularity — diff at the class level or the individual-target level; finer catches subtle redefinitions.
  • Change-gate strictness — advisory notice vs a hard pre-change approval before a mapping edit lands.
  • Evidence-expiry horizon — how old a witness may be before the diff reopens its target as provisional.

When it helps, and when it misleads

Its strength is closing the stale-coverage failure, where a map that was true a quarter ago is quietly false now; it makes drift visible at the moment of change rather than at the next external audit. Its failure mode is that it only catches what its triggers see — an untracked source change or an informal target redefinition slips straight through — and a noisy trigger set can bury genuine deltas in churn. The classic misuse is narrowing the triggers or the baseline so that inconvenient new obligations never register as changes at all.[1] The discipline that guards against it is to monitor source-side and target-side changes alike, and to hold the uncovered-by-default rule for every new or changed target.

How it implements the components

Target-Space Difference Review realizes the change-governance slice of the archetype — the components that keep coverage from silently decaying:

  • target_space_change_monitor — it watches for additions, splits, and redefinitions in the target set and flags each as a delta to re-cover.
  • mapping_change_control — it gates source retirements and relation edits through pre-change impact review, so a change cannot quietly orphan a covered target.
  • coverage_revalidation_cadence — it sets and drives the schedule on which the diff is re-run.

It does not build the original target list (that's Scenario Enumeration Workshop), decide who fixes each reopened gap (Uncovered-Target Triage), or re-execute the full witness recertification (Periodic Coverage Recertification).

References

[1] New obligations added without a matching witness are how a complete map silently becomes incomplete; treating each added or changed target as uncovered-until-proven is the standard guard — the coverage analogue of baselining a scope so that additions cannot be absorbed without review. ↩