Jump to text
On this page
Copy and source
Published source confirmed at last check

Source changed 2026-11-18 21:11:39 UTC · snapshot created 2026-11-18 21:12:35 UTC · last check 2026-11-18 22:10:10 UTC

A.19.CN:9 - Anti‑patterns (and the fix)

Anti‑patternSymptomWhy it hurtsFix (CN‑Spec slot)
Chartless number“Latency = 120”No unit/vantage → untestableFill cs_basis + chart
Normalization smugglingQuiet “per‑unit” normalisation mid‑streamTrend reversalDeclare UNM normalization references (NormalizationMethodId / NormalizationMethodInstanceId) + named invariants (see A.19.UNM)
Bridge-by-nameReusing equal labels under different schemesFalse comparabilityTest the needed F.9 predicate; cite a Bridge only when it obtains, then ground the separate receiving-use claim and reliance. An unsupported required correspondence blocks that use.
Free-hand averagingAn arithmetic mean is used without a model for the intended risk resultThe number can answer the wrong questionUse A.9 to select a justified law, bound or separate-input result; declare a Γ_fold only for the selected fold
CN‑frame sprawlOverlapping definitions obscure which comparison basis to useAvoidable reconstruction and maintenanceReuse a sufficient basis and retain differences that change the use. A registry supports maintained editions; use E.9 when a consequential decision needs a DRR.
Self-certification hidden by assignmentsA holder authors the basis and later certifies it under a new roleA required independent check still examines that holder’s own basisApply §4.2 to participation in the edition; separate authority and examination, and keep uncertified use explicit when certification is not required