Computed values and distributed hash-table splitting
Consider a situation like:
- Object A is stored on Node A
- Object B is stored on Node B, and has a schema which has a script that adds a computed value to Object A
Node A will never add the computed value that Object B generated to Object A, because it has no knowledge of Object B.
It would be preferable for Node A to have knowledge of that computed value; but that is currently impossible. Some protocol needs to be created to allow nodes to broadcast new computed values, as they do with new objects.