It's not the only way - casting one of the ints in your expression to a floating point type might be safer. If a mathematician was involved in writing your compiler, she might 'optimise out' a...