Re: [poe] Duty->Consequence

I forgot say that I agree with @fornaran and to thank her for the reference.
And I see now two interesting points in @riannella 's comment:

**1) Domain of failure**
I had only foreseen that a failure is activated:
* When a duty is not satisfied (and yet the permission's action acted)
* When a prohibition is violated
I have have totally missed:
* When a permission is exercised but the constraints not satisfied

In order to be general, I would include failure in this case, too. The red arrow should also depart from permission, the domain of the property should be "rule" instead and the range "rule". Even simpler model! Yet I would keep the three possibilities in the text definition.

**2) Conditional rules**
Let us consider P2, which I model below using version 2--although the problem is also common to version 1.
```
  ex:p1 odrl:prohibition [
    odrl:action ex:p2p ;
    odrl:assignee ex:me ;
    odrl:assigner ex:telco ;
    odrl:failure [
      a odrl:Duty ;
      odrl:assignee ex:telco ;
   odrl:action ex:report-police 
      ] 
  ]
```
If the RDF graph is queried about "how many duties there are", the failure/remedy duty here will be present. But we would like this will only to be true conditionally. Consequently, we have to add something to the Duty (in either version of the model) to indicate "only conditional". 
In order to count in-force duties, we can rely on sentences like "Duties that are independent, not object of any odrl:failure" but this may not be the best. The alternative would be to indicate explicitly that the duty is only conditionally in effect. I am neutral about either option.

-- 
GitHub Notification of comment by vroddon
Please view or discuss this issue at https://github.com/w3c/poe/issues/209#issuecomment-320932467 using your GitHub account

Received on Tuesday, 8 August 2017 11:50:34 UTC