If P is greater than zero, instead of storing the opposite value of P into D, the positive P can be stored. Doing so, however, the incrementation would need to be adjusted. Incrementing D by -1 would compensate for the difference.
Just another example of how the same algorithm does not require a unique code.