machine references vs identifiers

[dotnwat]

Consider the common situation in an implementation of a system where various resource identifiers (e.g. node_id) are stored and cached in ways that establish relationships between the resources.

In a P model of such a system where resources are modeled as machine, it seems common to have to choose between generating and using identifiers, and embedding references of machine type to represent the same relationships.

Are there advantages / disadvantages to either approach?

[ankushdesai]

Just to understand a bit better;

Is one of the reason to model resources as machine that you want to interact with the resources as well as these resources have state in them that change over time and you want to maintain that state in the resource state machines?

If yes, then I agree that modeling resource as a stateful state machine is the right choice and then instances of these machines represent resource instances.

[dotnwat]

Is one of the reason to model resources as machine that you want to interact with the resources as well as these resources have state in them that change over time and you want to maintain that state in the resource state machines?

exactly. you said it much better than myself.