prop vs class definitions

[VladimirAlexiev]

I'm not sure what form is the data underlying https://ns.mh1.eu/epcis/ but there are some tricky points on the way to using it to make the ontology (#205).

• It would be great if the same source can generate descriptive documentation for all of XML, JSON and RDF (ontology) but that may be very tricky.

See #207 for general principles of writing definitions. This issue is about the distinction between a property and its target class.

• 215 where bizTransaction should be removed and used as node URI; thus any documentation should be moved to the respective class bizTransaction (or bizTransactionDocument)

• Another example: although eventID is displayed as a prop at https://ns.mh1.eu/epcis/, it should be removed because it's just the URI of epcis:EPCISEvent, and its documentation moved to the class

[mgh128]

The underlying data for https://ns.mh1.eu/epcis/ , https://ns.mh1.eu/cbv/ and https://gs1.org/voc/ is a JSON-LD file for the ontology. In the case of the GS1 Web vocabulary, the JSON-LD file can also be accessed directly from this page: https://www.gs1.org/standards/gs1-smartsearch/1-6 . The intention is that direct access to the JSON-LD files for EPCIS and CBV will also be provided directly, as JSON-LD and RDF Turtle.

For https://ns.mh1.eu/epcis/ and https://ns.mh1.eu/cbv/ the rendering is all client-side using Vue.js. It's not currently relying on conversion of the JSON-LD to RDF triples. You can see the JSON-LD dataset and the JavaScript / Vue.js code if you view the HTML source code. These two tools are still under development.

I agree that it might be very tricky to include annotations to generate the XML documentation from this. The XSD files for EPCIS v1.2 are available at https://www.gs1.org/standards/epcis/artifacts/1-2 and the current validation schema drafts for EPCIS 2.0 (in XSD, JSON Schema and SHACL) are in this GitHub repository.

We appreciate the feedback, particularly from a Linked Data perspective and when you have formally joined the work group and signed the GS1 IP policy, we'll be able to formally consider your comments in further detail within the work group (of which you'll then be a member) - all your feedback so far appears to have significant merit.

Having said that, this is not entirely a 'green field' exercise and to some extent, we're held back by the XML/XSD data binding from previous versions of EPCIS that date back as far as 2007. We're definitely trying to modernise and make EPCIS data available as Linked Data but there might be some limitations on how far we can go, depending on how data has previously been expressed in XML - but the work group is open to constructive suggestions, especially for anything that does not result in major nomenclature differences between the XML and JSON/JSON-LD data bindings that are likely to confuse implementers or anyone migrating from the XML data binding to the JSON/JSON-LD data binding introduced in v2.0.

[VladimirAlexiev]

@mgh128

• I signed the 2 docs this morning.
• I agree with what you say re backward compatibility. No spurious changes should be made.
• I'll save the ontology JSONLD sources and convert to Turtle and put in source control in a folder "Ontology", so we have an explicit basis for discussion

[mgh128]

Thanks for doing those. We should receive confirmation soon from Eddy that they have been received and that there aren't any problems.

[CraigRe]

Still needs assessment to determine whether this can be done for 2.0, or needs deferral to 2.1.

Action for @mgh128 to experiment with extra annotations in Linked Data browser tools for handling aliases, ObjectProperty vs xsd:anyURI etc.

[VladimirAlexiev]

The last point is #238

[VladimirAlexiev]

@mgh128 The doc browser needs to use the two Meta-properties to express JSON keys for certain props and for class IRIs: https://github.com/gs1/EPCIS/issues/238#issuecomment-924131053