This makes it easier to implement each opcode, as the offset calculation and recovery of raw read bytes is internalized.
This only contains two instructions of which I know the correct output.