Activation Criteria for Attribute resolvers

[Tom Zeller]

> Do we want to think in terms of a
> similar predicate language to that used in filtering?

> With a stretch of imagination I could see repurposing all (or many) of the
> "natural policy rule" Match Functors (or at least those which only care
> about the AttributeRecipientContext), but I'd be worried about circularity
> of reference and also downright frightened of the Schema changes (how do you
> capture that "AttributeIssuerRegexpPolicyRule" is OK as an activation
> criteria, but that "NumOfAttributeValuesPolicyRule" isn't).

The initial thought of genericizing match functors with attribute
resolution activation criteria frightens me too, I am okay with
keeping functors and criteria separate for now.