USDM v4 · overlay graph · draft, not a CDISC artifact
usdm-overlay reads the USDM v4 model as rendered by usdm-rdf and classifies every attribute and class by the kind of content it carries. Every reading names its evidence. This page lets you browse those readings and the findings they surface.
How it fits
Readings of the model: content kind per attribute, role per class, findings with tests. Points at usdm-rdf identifiers and never changes them.
USDM v4 rendered mechanically in RDF/OWL, with a stable identifier for every class and attribute.
The published model, controlled terminology and example studies (DDF-RA v4.0.0).
Layer 1 · content kind per attribute
Wording people write and read: rationale, conditions, classifications, narrative.
Strings that name, identify, cite or encode: identifiers, codes, ISO 8601 values, address parts.
Layer 2 · role per class
A role describes what the class itself holds, counting its own and inherited attributes, without housekeeping, naming and notes. Pick a role, or search, then open a class.
Layer 3 · findings, each with a test
Each finding is rerun on every build. Its standing is separate from its evidence: a gap the model states can still be a deliberate design choice.