A.19 already declares CharacteristicSpace, but source-set and publication-form semantics still need a separate substrate declaration.
Precision vs over-typing
The line needs explicit ref positions, an explicit ref-to-ref relation kind, and explicit relation posture, but it should not fabricate composition, derivation, metrics, or transition qualifier when the case does not need them.
Reuse vs semantic collapse
DescriptorMapRef, DistanceDefRef, OutcomeMapRef, or BridgeDistortionNote are useful qualifiers, but they must not silently become the whole substrate.
User readability vs architectural honesty
Cold readers need a first-minute explanation, while specialist readers still need exact boundaries and docking rules.
Interpretive views vs substrate core
Atlas or interpretive-view lines can be valuable, but they should remain optional derived help rather than the default meaning of the substrate.
Uncertainty honesty vs fake closure
Many current lines use learned, adaptive, unstructured, or distribution-valued spaces or relations; the pattern must expose that posture without pretending the heaviest qualification posture is already settled.