PRE #2 — code lists baked into schema, money flattened to literals

The second defect cluster: reference data modelled as closed schema enums (three different ways), and monetary amounts stripped of currency and time.

The 5 defects

  1. 1Schema-frozen enums. Status values are display strings inside gist:Category classes: no IRIs to reference, no skos:notation for the business codes, and every value change is an ontology release. The project standard (rebuttal-codelist-skos) says: code lists are reference data — SKOS schemes checked by SHACL sh:in.
  2. 2A process in the type list. “Subrogation” is a recovery process; putting it beside “First Party Property” makes type statistics meaningless. (CLM-6, CLM-12)
  3. 3Two patterns in one delivery. pandc:DOClaimType already uses skos:Concept — consumers need two query shapes for the same kind of value. (CLM-6)
  4. 4Money flattened. Bare decimals with no currency or valuation date; the stored incurred figure can silently contradict paid + reserve. (CLM-7, CLM-8)
  5. 5Payments floating free. Optional claim link + an empty payment-type class: indemnity/expense splits cannot reconcile. (CLM-11)