[Application-profiles-ig] Shapes and cardinality

Thu Oct 1 07:12:23 BST 2020

On 2020-09-30 06:19, Phil Barker wrote: >> However, I do think I addressed your requirement in the rest of my>> response (see below). Or am I still getting the question wrong?>>Well, sort of addressed as "won't do" (unless I'm misreading).Hi Phil, To be clear: - I agree that your requirement to record this information can be met with columns for cardinality on shapes. - However, I object to having the existing Opt/Repeat columns do double-duty for cardinality either on a shape or on a statement and propose that we clarify that the existing columns are for cardinality on statements. - On the other hand, columns specifically for shape cardinality could be added to a DCAP CSV -- even if not specified in the core model -- without breaking anything. In general, I think we should invite people (with appropriate caveats) to add application-specific columns to a CSV as they see fit. - I note with interest that validating for shape cardinality over an entire graph is recognized as a use case in the ShEx community, and there are various ways this could be addressed by extending ShEx (or its Shape Map specification). However, I take the fact that this has not previously been much discussed as a potential feature for ShEx as a lack of strong demand. - I would therefore not favor including shape cardinality columns in our core model. - I would, however, have no objection to including this in a section of the spec (or primer) about how to extend the CSV model if our group agrees. >Which is fine. I'm still on the path I mentioned>earlier in this thread, of using the pattern of one>表项属性(符合what>we've done here) and other sheets for terms that are>classes and for the profile as a whole.I think I understand your requirement now and do not question that this is a reasonable way to do things. Tom -- Tom Baker <tom at tombaker.org>