N digit signed-magnitude number is at most $-2^{n-1}+1 < = sA < = 2^{n-1}-1$
first digit 1 is negative, 0 is positive.
There is two zero representations. 100000. or 000000.
We want to solve the problem that the magnitude is separated with its sign.
We want just simple bit-wise addition possible.
numerically integrate the sign and its magnitude.