[Odrl-version2] Core Model: Roles, Parties and Assets

[Francis Cave]

Here is a set of related issues that have arisen during discussions among
the EDItEUR team:

 

The latest draft of the Core Model allows a Permission or Prohibition to
have multiple associated Parties. It is mandatory to have at least one
associated Party. The introduction of the Role entity raises a number of
questions.

 

1. The new Role entity appears to be optional, although neither Figure 2.1
nor the text of 2.4 is clear on this point. What is the intention here? The
dotted lines in Figure 2.1 suggest that the precise relationship of the new
Role entity with the existing entities is not yet settled - or is this a UML
convention with which we are not familiar?

 

2. If it is mandatory for a Permission / Prohibition to have at least one
associated Party, does this contradict Figure 3.1 (Set), in which apparently
a Policy does not have a Party? The model for a Ticket also seem to
contradict 2.1.

 

3. Given the rationale for allowing multiple Parties associated with a
Permission / Prohibition, by analogy would it not make sense to allow
multiple Assets to be associated with a Permission / Prohibition, with the
Role entity used to define the association between Permission / Prohibition
and Asset? This would make it possible to express such things as "it is
permitted/prohibited to perform Action X on Asset A, where Action X involves
Asset B". The fact that this would also improve the symmetry of Figure 2.1
is a bonus, but probably not significant.

 

4. The Role entity currently has two properties: 'function' and 'scope'. The
first if these is clearly required. The second appears to be descriptive of
the Party and not of the Role, and we wonder whether we have misunderstood
its purpose. We are fairly clear what is meant by "individual": an
identified individual person or corporate entity. But we are not completely
sure what is meant by "group". The semantics of "all" (one of the additional
values for 'scope' added to the latest draft of the Common Vocabulary) seems
to suggest that the intention is that "group" means "any member of the
identified group", while "all" means "all members of the identified group"
(i.e. the whole group). Have we understood this correctly?

 

Francis Cave

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://odrl.net/pipermail/odrl-version2_odrl.net/attachments/20101105/f3455f89/attachment.html>