Annotation`sh:severity`

[afs]

The domain of sh:severity is sh:Shape

SHACL NS

sh:severity
	a rdf:Property ;
	rdfs:label "severity"@en ;
	rdfs:comment "Defines the severity that validation results produced by a shape must have. Defaults to sh:Violation."@en ;
	rdfs:domain sh:Shape ;
	rdfs:range sh:Severity ;
	rdfs:isDefinedBy sh: .

And this is in agreement with the diagram in [https://www.w3.org/TR/shacl12-core/#constraints-section](3. Shapes and Constraints).

In SHACL 1.2 the section is "3.1.4 Declaring the Severity of a Shape or Constraint"

Shapes can specify one value for the property sh:severity in the shapes graph. Each value of sh:severity is an IRI.

then the table #646. As a grey box, it is incomplete as a syntax rule. (aside: dark mode issue here)

Only later does it say:

Let T be the set of triples that represent a constraint in a shape. A shapes graph can specify at most one value for the property sh:severity in the reifiers of the triples in T.

The syntax rule should be complete and include both in the same box.

From TopQuadrant/shacl#199

Also - what happens if sh:message is in the annotation and in the SPARQL constraint?

[HolgerKnublauch]

Ok, so I guess we could remove the rdfs:domain here to also allow the reification case.

On the sh:message I would add a bullet point to the sh:resultMessage in SHACL-SPARQL to clarify that it will fall back to the defaults.