Gary O'Neall

> please keep in mind that the human-editable markdown files were not meant to be used for anything else besides the spec-parser. @zvr - that is exactly my point -...

> @goneall @zvr: I'd ask you both to look at [Relationship](https://github.com/spdx/spdx-3-serialization-prototype-playground/blob/b60b3de89af68ba94b53e49ed8f83baa67f1d8b6/model/generated_20230705T100948Z/Core/Relationship.md). > > This is the logical value of a relationship element. which has start and end DateTimes. The logical...

I spent some time trying to get a definitive answer on if you can subclass data types in RDF. It looks like I was wrong - you can subclass datatypes....

> It looks like I was wrong - you can subclass datatypes. Based on this, I withdraw my objection and it looks like we are property treating DateTime property correctly....

I found this [W3C note](https://www.w3.org/TR/swbp-xsch-datatypes/) that provides some context on using data types. For reference, here's the [W3C description of data types](https://www.w3.org/TR/rdf12-concepts/#section-Datatypes) in the concepts document. My reading of the...

I think for RDF, my current opinion is we "restrict" our use to data types to only be pre-defined XSD schema datatypes and use SHACL to further restrict the property....

Back to the original PR - in RDF, the `SpdxId` will always have to be the URI of the object itself for the graph to work. It seems a bit...

@davaya - My comment above is rather specific to RDF and the Ontology and Schema generated by the spec parser. In it's current form, we are duplicating the ID -...

Since this is targeted for RC2 - @zvr @davaya - we should discuss within the next couple weeks and come to a conclusion

> We have yet to hear any specific harms, costs, or disadvantages. I'll register one disadvantage. It is specific to RDF serializations. We currently make the `spdxId` property the ID...