- From: Joshua Cornejo <josh@marketdata.md>
- Date: Thu, 13 Jun 2024 08:32:40 +0100
- To: "public-odrl@w3.org Group" <public-odrl@w3.org>
- Message-ID: <5EFC269B-2B4E-4707-942F-6302EBF4F6D1@marketdata.md>
I am looking for a constraint-based model for the tuples of [action -> parties] to prevent the following 2 situations:
As your catalogue of policies grows (policies + roles + departments + parties), it will create unnecessary rules that need to be managed and tracked in sets (extra admin + more chance of errors).
If I want to map to the control layer in REGO (OPA) or CEL (FGA), I have N ODRL rules = 1 REGU/CEL rule (N = number of parties) – now my ‘mapper’ has to be extra smart to bound to the output (complex), or I will have a larger (unnecessary) ruleset to test (runtime inefficient).
Regards,
___________________________________
Joshua Cornejo
marketdata
embed open standards
across your supply chain
From: Renato Iannella <r@iannel.la>
Date: Thursday 13 June 2024 at 04:04
To: "public-odrl@w3.org Group" <public-odrl@w3.org>
Subject: Re: constraining parties and leftOperands
Resent-From: <public-odrl@w3.org>
Resent-Date: Thu, 13 Jun 2024 03:04:29 +0000
I may have missed some complexity, but would this work:
ex:policy001 a odrl:Agreement ;
odrl:assigner <http://example.com/party:owner> ;
odrl:target <http://example.com/asset:1> ;
odrl:permission [
odrl:assignee <http://example.com/party:user> ;
odrl:action odrl:play
] ;
odrl:permission [
odrl:assignee <http://example.com/party:administrator> ;
odrl:action odrl:transform
] .
Cheers - R
On 12 Jun 2024, at 18:30, Joshua Cornejo <josh@marketdata.md> wrote:
Thanks,
Discarding that operand, the original question remains unanswered.
How can I model these common restrictions with the current LeftOperand taxonomy:
ex:User (Renato) can odrl:play an Asset (read) , but ex:Administrator (Victor) can odrl:transform (read/write)?
<image001.png>
If odrl:industry wasn’t so narrowly defined, it could fit the purpose.
Regards,
Received on Thursday, 13 June 2024 07:32:50 UTC