This an overloaded construct, which serves multiple purposes:

1. it is used by ParserIdentifier as a return value, serving all callers who wish to parse a request URL's resource identifier part
2. is is used in NormalizedNodeContext as the context in which the result should be interpreted
3. it is used intermittenly along the path between 1) and 2)

At the end of the day, with current state of refactoring, we are completely in control of the call sites that interact with 1) and are in a position resolution rules. For example, in POST we are able to say that 'parser this identifier as a datastore reference or a invocation reference'. These two results should be backed by separate records.

For part 2), we want to inline the inforation, which really should boil down to an Inference, into NormalizedNodeContext. This part may be plagued by legacy, so needs to be looked at carefully.

Transitions between the two should be easy to model and simple to translate, so 3) should be simple after the first two are dealt with.