In this case, that was deliberate. I was using it to show the exact decimal value of the floating point approximations to decimal numbers, not to do decimal arithmetic.