The move operation simply relocates a node (and its descendants) to a different location, but perhaps more importantly moved nodes keep identifiers. References to nodes that are moved will remain valid after the move.

Like copy, there are several possibilities with respect to moves that involve federated content:

• The source content includes federated nodes, but the target is not federated. This is essentially copying external content into the repository. This makes no sense since we cannot keep the identifiers the same, so we should not support this and throw an exception. It could be replaced by a copy operation and a remove of the source, but any references would not be migrated.

• The source content does not include federated nodes, but the target location itself is federated. This is essentially moving internal (non-federated) content into an external source. Again, identifiers of external nodes are simply different than internal nodes, so we can't keep them the same. We should also throw an exception in this case.

• The source is federated and the target is federated. In case the move operation occurs within a single (the same) source and the connector is writable, we should support this. If the move happens across sources an exception should be raised, since we can't keep identifiers the same.