Using unsigned data types is acceptable as signed operations are much more complex than unsigned ones (e.g., shifts, divisions and bitfield operations). But it is a bad idea to use unsigned types just because a value is always larger or equal to zero, and because the type can hold a larger positive number.