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

Source changed 2026-11-18 22:51:26 UTC · snapshot created 2026-11-18 22:54:45 UTC · last check 2026-11-18 23:05:20 UTC

A.19.UNM:8 - Common Anti‑Patterns and How to Avoid Them

  1. Hidden normalization inside scoring or selection Avoid this by using CN_Spec.comparability.mode and applying UNM.

  2. “NCV ⇒ indicator” shortcut Avoid by treating indicatorization as UINDM policy, not a byproduct of normalization.

  3. “We normalized” without declaring invariants Avoid by naming the domain, transformation, preserved invariants and lost distinctions; supply a class or congruence claim only under its additional conditions.

  4. Reusing a normalized value after its basis changed Recover the bearer, method and CN-Spec editions, comparison basis, evidence, intended use and scope/window. Recheck the conditions that changed, including any relation required by the new use.

  5. Choosing a representative implicitly Avoid by either keeping quotient objects abstract or declaring NormalizationFix.

  6. Treating a generic mapping word as a specialized relation claim State the normalization’s operands, rule, invariants and loss. Test any specialized Map or F.9 Bridge claim separately under its defining pattern.

  7. Treating UNM outputs as comparable beyond their declared bearer, basis, scope/window, or reference plane Avoid by keeping comparison local to the recorded premises. Where a conclusion depends on another source-local meaning, bearer kind, or plane, cite the obtaining relation and its loss; otherwise constitute a new normalization result or fail closed.

  8. Re-authoring method, basis, or evidence anchors downstream Avoid by citing the editioned method, basis, and evidence anchors as refs; a downstream pattern neither rewrites them nor replaces them with a generic registry.