TODO --- this is unDRYly like the above functions and could be coalesced with them by making the types of nodes to be captured parameters of the function.

We want to extract the prefixes used in a 'given' node.

This works with the following function to find dependencies in a 'given' block.

We want to extract the variables used in a given node.

This exists because we have too many ways of representing types.

The base Node interface.