>We didn't state it in the specification? But I guess we are still on
the
>page on this one, anyway.
I hope we did not. Unlike specifics of the functions behaviour (and
properties of the expression references for that matter), virtual
context is not ours to regulate, I believe.
>I thought the reasons we adopted unordered bag functions was because of
>the shouldn't rely on getting the answers back in a specific order for
a
>query evaluation. That is, one PDP may get them back in a different
order
>than another PDP.
I thought it is not only about different PDPs. Nothing explicitly
prohibits a single PDP to be distributed and retrieve data from context
for single policy evaluation from different "directions". And get a
different answer for a different rule. Normally you would not want such
behavior, but that's not always about common sense. I have a working
example of a system where this is not the case - context handler is out
of control of the evaluation and very well may change between individual
rule evaluations (yuck, but true. And efficient)
>I think the situation you mention, within the same PDP during the same
>evaluation is merely an implementation that doesn't violate the
>constraints and is acceptable.
Indeed. The problem with that was we could not write a policy that
REQUIRES the same expression to be evaluated to the same value within
the single instance of policy evaluation. With expression references
that can be solved - without imposing unduly requirements on policies
that do not require that.
> Agreed.
Same.
Daniel.
