SHACL packaging

[nicholascar]

• introduces a section on the packaging of SHACL to the Profiling deliverable

See this rendered here:

• https://raw.githack.com/w3c/data-shapes/shacl-packaging/shacl12-profiling/index.html

[TallTed]

Getting there...

[TallTed]

Multiple instances of —

@nicholascar nicholascar deleted a comment from TallTed 4 days ago

I don't know what to do with this. It doesn't give me a way to see the deleted comments, so I have no way of modifying my future comments, nor can I defend these comments. Please advise!

[scotthenninger]

Changes look good to me

[mgberg]

@nicholascar out of curiosity, for your use case, would you expect each group of shapes from your shape library to be represented as an instance of owl:Ontology?

[nicholascar]

Multiple instances of —

@nicholascar nicholascar deleted a comment from TallTed 4 days ago

I don't know what to do with this. It doesn't give me a way to see the deleted comments, so I have no way of modifying my future comments, nor can I defend these comments. Please advise!

Sorry Ted. I tried to incorporate the suggestions from every one of your comments. I merged what I could right in this GitHub UI and then, where there were conflicts due to me editing after your comments, I tried to make the change manually and noted that I had done that.

I tried to be consistent with "instances of" and so on.

I think the comment deletion is an outcome of me indicating I thought the suggestion was incorporated, rather than me saying "no". There were no instances of your comments that I didn't agree with and try to incorporate, except where I had already made an equivalent change. None I thought were wrong or not valuable.

[nicholascar]

@nicholascar out of curiosity, for your use case, would you expect each group of shapes from your shape library to be represented as an instance of owl:Ontology?

Yes, owl:Ontology.

I used to want special container classes for things - makes them easy to find in a DB - but now I go in for containers being generic so I don't have a profusion of container types and find them by one of these means:

• seeking downwards from an upper container, e.g. a dedicated dcat:Catalog of owl:Ontology instances for SHACL
• duck typing: "find me all the ontologies that contain (rdfs:member) SHACL Elements (and not owl:Class etc)
• role-based: find me all the resource members of prof:Profile instances that conform to SHACL

In a tightly managed collection, I use the first method but the second would requires less management and be more robust if I could be guaranteed that all the SHACL was presented as per this Spec's recommendations!

The third method requires a Profile hierarchy according to PROF, which of course I use but which adds another index to the SHACL info, just like making a DCAT catalogue of things.