– DRAFT –
Data Shapes WG weekly teleconference

Announcements

Discussion

<AndyS> w3c/data-shapes#408

mgberg: "Add sh:nodeByExpression Constraint Component"
… provide a node expression to compute the node shapes to validate against
… just using sh:node + expression does not work (because of blank nodes)
… example UC -- data cube vocabulary
… currently no good way to apply the constraints the focus nodes (current, 1.0 - it's clunky)

elianaP: questions?
… (silence)

<AndyS> w3c/data-shapes#386 -- "Added script to sync Turtle & JSON-LD snippets"

script + changes required to allow the script to work

AndyS: someone mentioned doing the same for SHACL compact

bergos: Some work may be required to verify that the generated SHACL conforms to the examples

AndyS: SHACL-C generates SHACL RDF triples that are used for validation- same triples under the hood

<AndyS> jeswr -- w3c/data-shapes#419

AndyS: will follow up on the SHACL-C issue

<AndyS> w3c/data-shapes#409 -- "add severity levels that are not violations"

AndyS: adds new severity levels, some are violating, some are not
… YoucTagh: Notes there are no examples of some of the severity levels, e.g. a developer use case
… AndyS: suggests extending the test suite with extra examples, YoucTagh mentions that although useful users would likely not use the test suite

<AndyS> w3c/data-shapes#410 -- "add support for sh:TripleTerm to sh:nodeKind and allow lists"

AndyS: ajnelson-nist: Suggests adding example of conforming debug result in the example
… open question- deprecate the OR variations of node kinds to support lists? How would that happen?
… HolgerK: Downsides to duplicate/redundant syntax, use cases for sh:TripleTerm or other types seem unlikely. Suggests not using OR lists and using prior convention

<caribou> +1 to status quo unless we gather new use cases that really needs lists

AndyS: ajnelson-nist: Agrees with above sentiment
… bergos: at this point, the current clear use case is that RDF 1.2 triple terms are constructed as expected

<ajnelson-nist> Possible example - ?x ex:supports ?y - ?y a hypothetical statement, ex:supports in some ontology about citations?

ajnelson-nist: Potentially could use CiTO (https://sparontologies.github.io/cito/current/cito.html) as inspiration for the use case

<ajnelson-nist> ( Disclaimer: Participation by NIST in the creation of the documentation of mentioned software is not intended to imply a recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to imply that any specific software is necessarily the best available for the purpose. )

bergos: : Some additional complexity in the engine with OR, but may be simpler for users to pick one or more options from a list

HolgerK: Some users will have algorithms that will need replacing, e.g. determining whether a shape represents an object property or a datatype property

ajnelson-nist suggests redefining current IRIs as OR lists

HolgerK: mentions that won't work as the sh:or predicate would have to be used instead
… lists could include e.g. sh:BlankNodeOrIRI in the lists which could also add some additional confusion

<YoucTagh> w3c/data-shapes#390

<ajnelson-nist> w3c/data-shapes#362

HolgerK: PR with changes to test case namespaces has not been merged, but other PRs have been merged that modify test cases