ODRL state of the world representation

[AndreaCimminoArriaga]

The actual ODRL does not promote a formal representation for the state of the world (related to issue #61). Some thoughts about this issue:

• What is the definition of the state of the world and who provides it during the evaluation? other standards like XACML just define this concept as data values to define the additional information that is required in order to evaluate a policy. Allow me to assume this rather generic definition for the next points.
• In order to evaluate a policy, certain data will be needed. This data may come from the system that is evaluating the policy (e.g., the current date and time), a requester (who may provide certain data needed in the policy), or an external system.
• The nature of the data needed for a policy to be evaluated may be diverse in each policy (rdf, string values, jsons, etc.).
• The state of the world, regardless its formal representation, should be hidden and separated from the policy otherwise anyone that could read the policy would know the conditions that must be met.

In other words, an ODRL policy before evaluation should contain missing parts and, at the evaluation time, fill them with the state of the world data (regardless synchronous or asynchronous evaluation). On this sense, ODRL operands or operators can be suitable to refer to certain data that will be present only during evaluation (e.g., odrl:dateTime). However, note that these operators may not be enough if they stand for requester's data.

I think, IMHO, this could be clarified in the text of the formal semantics document. In particular, a closer definition of the state of the world, who can provide it, and maybe the fact that a policy should not display it (in the case it could) until it is evaluated.

In our research, to address this issue, we defined policies B1, B2, and C that rely on HTML templating languages to hide certain data values from the state of the world.