Re: [poe] Model clarifications

4. Maybe while reading the next version at once I will still feel uneasy with 'composition' itself. But right now it seems a better wording, anyway!

5. Maybe a policy that (possibly by inheritance) asks the assignee to attribute the creator and at the same time to anonymize the asset?
Or a policy that declares inheritance from CC-BY but requires compensation for commercial use?
They will all look a bit absurd of course, but I guess it's the essence of conflict :-)

6/7. Maybe there's some confusion? #6 was already ok for me. #7 was the hard one.
Adding @type and using constraints brings clarity. It still doesn't solve the essential issue of URI hijacking, though. The problem lies in the resources created/re-used in the constraints, not their types or their statements. As long as different statements are made about the same URI, then confusing arises and the same issues of 'conflating triples' will happen, as I've put in example (just replace scope statements with statements about constraints and types in my example above).

11. (this is both discussions on data types and right operands URIs are relevant)
data type: https://w3c.github.io/poe/vocab/#term-dataType looks much better, but I still don't get why it's a MAY not a must. This is opening the door to several (quite different) representations of the same information in RDF. And as a matter of fact example 14 (and maybe others) still don't use the right RDF pattern. Do you plan to update them?

righOperand vs rightOperandReference: OK I think I get it now. I would still urge you to go one step further in clarifying this modeling, which can be confusing to the reader: what I can think of now:
- declaring rightOperand a Datatype property (in the OWL sense).
-  saying somewhere that even when rightOperand is used with URLs, these are to be considered as strings
- changing the definition "The right operand in a constraint expression" at https://w3c.github.io/poe/vocab/#term-rightOperand (and maybe elsewhere where the property is mentioned?). Right now it reads too generic - as if it was also covering the cases covered by rightOperandReference


-- 
GitHub Notification of comment by aisaac
Please view or discuss this issue at https://github.com/w3c/poe/issues/162#issuecomment-306453394 using your GitHub account

Received on Tuesday, 6 June 2017 11:05:01 UTC