SHACL separates the shapes graph from the data graph. NodeShapes select targets; PropertyShapes define path, cardinality, datatype, class, and other constraints. Validation produces a results graph.

1. 1

   Define contract

   Define contract

   

   Goal

   Produce constraint backlog so the core question “Which RDF structure must hold for the use case, and how is each violation reported?” can be answered with stronger support.

   Inputs for this step

   • RDF use cases
   • valid and invalid sample data
   • namespace and severity policy

   Actions

   • Define use case, target, must/should rules, and severity.
   • Record source, rationale, and open uncertainty directly in constraint backlog.

   Output

   constraint backlog

   constraint backlog · source · uncertainty · next decision

   Quality check

   Is every rule justified by the domain?

   Pitfall

   Shapes copy accidental patterns from sample data instead of justified domain constraints. Document use case and constraint rationale, write positive and negative cases, and set severity deliberately.

   Timing: 10–20 minutes

2. 2

   Author shapes

   Author shapes

   

   Goal

   Produce shapes graph so the core question “Which RDF structure must hold for the use case, and how is each violation reported?” can be answered with stronger support.

   Inputs for this step

   • constraint backlog
   • valid and invalid sample data
   • namespace and severity policy

   Actions

   • Write NodeShapes and PropertyShapes with Core components.
   • Record source, rationale, and open uncertainty directly in shapes graph.

   Output

   shapes graph

   shapes graph · source · uncertainty · next decision

   Quality check

   Are paths, counts, datatypes, and classes correct?

   Pitfall

   Shapes copy accidental patterns from sample data instead of justified domain constraints. Document use case and constraint rationale, write positive and negative cases, and set severity deliberately.

   Timing: 10–20 minutes

3. 3

   Build fixtures

   Build fixtures

   

   Goal

   Produce RDF fixtures so the core question “Which RDF structure must hold for the use case, and how is each violation reported?” can be answered with stronger support.

   Inputs for this step

   • shapes graph
   • valid and invalid sample data
   • namespace and severity policy

   Actions

   • Create valid, invalid, and boundary data.
   • Record source, rationale, and open uncertainty directly in RDF fixtures.

   Output

   RDF fixtures

   RDF fixtures · source · uncertainty · next decision

   Quality check

   Does each negative case violate the intended rule?

   Pitfall

   Shapes copy accidental patterns from sample data instead of justified domain constraints. Document use case and constraint rationale, write positive and negative cases, and set severity deliberately.

   Timing: 10–20 minutes

4. 4

   Validate

   Validate

   

   Goal

   Produce validation suite so the core question “Which RDF structure must hold for the use case, and how is each violation reported?” can be answered with stronger support.

   Inputs for this step

   • RDF fixtures
   • valid and invalid sample data
   • namespace and severity policy

   Actions

   • Run a SHACL processor and inspect the results graph.
   • Record source, rationale, and open uncertainty directly in validation suite.

   Output

   validation suite

   validation suite · source · uncertainty · next decision

   Quality check

   Does every result include focus node, path, and constraint?

   Pitfall

   Shapes copy accidental patterns from sample data instead of justified domain constraints. Document use case and constraint rationale, write positive and negative cases, and set severity deliberately.

   Timing: 10–20 minutes