Extending expressiveness for evaluations

[joshcornejo]

Considering the input triple <subject, action, resource> and to prevent the unnecessary repetition of rules to cover different levels whilst performing the same conditions (e.g. dcat:Catalog -> dcat:Catalog -> dcat:Dataset -> etc or hierarchical party collection - which are less common), a few keywords and rules could be introduced:

• new:any - placeholder for understanding: new:any a odrl:PartyCollection, odrl:AssetCollection .
• new:resource leftOperand that refers to the resource in the input triple
• new:subject leftOperand that refers to the subject in the input triple (could also be odrl:recipient)
• Use of new:any must always be refined with a odrl:hasPart or odrl:isPartOf.

fragment:

{
  "target": "new:any",
  "refinement": {    
      "leftOperand": "new:resource",
      "operator": "hasPart", 
      "rightOperand": "someAssetCollection"
      }
}

pseudo-code shortened version:

            target any { refinement idForAsset { resource hasPart someAssetCollection } 
            assignee any { refinement idForParty { subject hasPart somePartyCollection } 

            target any { refinement idForAsset { resource isPartOf someAssetCollection } 
            assignee any { refinement idForParty { subject isPartOf somePartyCollection } 
}

(note - this would be relatively simple to achieve in *BNF / context free grammars)