Re: [poe] Split LeftOperands into constraints and refinements

I've raised two concerns about the use of LeftOperand in my comments:

1. My concerns of constraint Constraint vs refinement Constraint are in fact only about the LeftOperand instances.
The IM specs regarding the use of LeftOperand instances are ok, I see no need to change them.
I suggested to advise users which of the LeftOperand instances listed in the Common Vocabulary should be used with constraint or refinement - or both. And I pointed at the fact that about 20 of theses LeftOperand instances are related to the Action of a rule which will raise the question by users: create a Rule constraint or an Action refinement with that LeftOperand?
Note: if there is no problem to apply a Constraint with such an action-related LeftOperand as constraint or as refinement, why do we have these two variants at all?
2. I point at an issue raised by @simonstey that if a refinement Constraint is satisfied is very hard to evaluate. The only realistic evaluation is if an action is exercised WITH all its refinements in a "satisfied" state. If this view is shared by the WG the narrative about that should be edited.  E.g. in IM 2.5.4 ... "To meet this condition of narrower semantics for the Action, all of the Constraints/Logical Constraints referenced by the refinement property MUST **be used as generating a satisfied state**." 
Note: by my experience with the implementation of IPTC standards I don't support a heavy use of "black boxes" as this has a negative impact on interoperability. And this should not be the case for a standard used in the area of rights.


-- 
GitHub Notification of comment by nitmws
Please view or discuss this issue at https://github.com/w3c/poe/issues/282#issuecomment-348894873 using your GitHub account

Received on Monday, 4 December 2017 08:39:23 UTC