PRFO1

This objection is a repeat of the previously raised CRFO2.

This is a formal objection to the fact that "SHACL implementations are not required to have a mode that checks if a shapes graph is ill-formed and provide access to the results of that check - absence of a requirement for such mode to be provided by all implementations impacts interoperability.".

Submitted by: Peter Patel-Schneider

Description of the issue

SHACL implementations have to comply with requirements for treating well formed shapes. In case of ill formed shapes, behavior is left to implementations. For example, one implementation could ignore the ill formed shapes in a shapes graph and only process well formed shapes. Another implementation may refuse to process the entire graph.

Syntax checking mode that tells a user whether a shapes graph is well formed is described as optional in the SHACL specification. As a result, SHACL shapes creators can not rely on a fact that a given SHACL implementation would tell them whether their shape graphs are well formed. Thus, they would not know whether data validation against their shapes graphs would work inter-operably across all SHACL implementations.

The submitter believes that SHACL meta-shapes graph created at CR is inadequate to address this problem because it can't detect errors in regular expressions such as:

 ex:s1 a sh:PropertyShape ;
   sh:targetClass ex:C ;
   sh:path ex:p ;
   sh:datatype xsd:string ;
   sh:pattern "abcde(fg" .

Nor can the meta-shapes graph check for circular lists and paths or recursive shapes.

Submitter believes that the only way to address this issue is to require all SHACL implementations to have a mode (like strict mode in many compilers) where they signal whenever they are given a shapes graph that contains ill-formed SHACL constructs or SHACL constructs that they do not handle.

Relevant information

E-mail

To ensure 100% interoperability across implementations, users need to check their shapes graphs before they are sent to the engine. SHACL meta-shapes graph will detect the majority of issues, but not all issues.

WG Response

WG feels that PRFO1 was addressed sufficiently in its prior incarnation as CRFO2