INT DX, J, AX;
....
J = DX <0? -1: 0;
AX = (DX ^ j) - j;
Why write this AX assignment statement? Test proves to be more conventional
AX = AX <0? -ax: ax;
Slowly
