Analysis Dimensions
The workbench analyzes SHACL-based digital passport model specifications without requiring representative instance datasets.
Structural Size And Complexity
Structural metrics describe the scale and shape of a model composition.
Examples include:
- number of node shapes
- number of property shapes
- number of target declarations
- number of closed shapes
- number of datatype-constrained properties
- number of object-reference properties
These metrics help distinguish a compact model slice from a broader composed ecosystem.
Explicitness And Constraint Readiness
Explicitness metrics describe how much of the model is directly constrained in SHACL.
Examples include:
- typed-property share
- cardinality-bounded-property share
- open-property share
- constraint density
These metrics are useful when comparing whether one model expresses constraints more directly than another.
Structural Interoperability
Interoperability metrics describe relationships between modules in a composed profile.
Examples include:
- cross-module reference count
- cross-module reference share
- shared-vocabulary overlap count
- shared-vocabulary overlap ratio
Release 1 treats each top-level models entry in a composition profile as one module boundary.
Maintainability Signals
Maintainability findings are directly detectable issues or candidates from the SHACL graph.
Examples include:
- contradictions
- dangling references
- redundancy candidates
These findings are intentionally conservative. They are meant to be inspectable signals, not hidden semantic judgments.
Use-Case Coverage
Coverage analysis checks whether a declared use case can be represented from the composed SHACL model.
Use cases are declared as:
- required information items
- required joins or concept links
Coverage classes are:
representablepartially_representablenot_representableindeterminate
The result is SHACL-only. It does not require instance data.
Model Comparison
Comparison runs on two assessment result documents with the same comparison_scope_label.
It reports:
- metric values for both sides
- deltas
- normalized ranked observations
- optional alignment-aware comparison when an analyst-authored alignment file is supplied
This supports both version comparison and cross-ecosystem comparison.
Prioritization
Prioritization consumes previous result documents and ranks follow-up targets.
Signals can come from:
- coverage gaps
- maintainability findings
- directional comparison deltas
- alignment gaps
The prioritization layer is rule-based and explainable by design.
Result Interpretation
The summarize operation provides a lightweight interpretation layer over existing result documents.
It does not add new analysis. It extracts a concise headline, key points, and follow-up questions from already computed results.