A.19.UINDM:6 - Bias-Annotation (informative)
-
Gov (governance). Bias toward explicit policy surfaces (
IndicatorChoicePolicyRef, edition pins, auditable outcomes) rather than tacit “expert choice.” Risk: perceived extra work. Mitigation: keep the mechanism minimal (selection‑only) and push method detail into wiring modules. -
Arch (architecture). Bias toward stable interfaces: SlotKind tokens come from the suite lexicon and evidence gates are explicit inputs. Risk: reduced “quick hacks.” Mitigation: allow
⊑⁺extensions for additional outputs (e.g., justification) without mutating the kernel signature. -
Onto/Epist. Bias toward a distinction between “measurable characteristic” and “indicator under policy.” Risk: teams accustomed to “everything measurable is an indicator” may resist. Mitigation: embed this as a LawSet clause (“No implicit NCV⇒indicator”).
-
Prag (pragmatics). Bias toward fail‑closed guards and traceability under uncertainty. Risk: more
abstain/degradeoutcomes early. Mitigation: coupledegradewith explicit downstream behaviors (policy‑bound) rather than silent coercions.