Citation:
"When we can, we prefer to limit ourselves to a few data types. That can help minimize confusion. For example, by not using float in a program, but only double, we eliminate the possibility of double-to-float conversion problems. In fact, we prefer to limit our use to int, double and complex for computation, char for characters, and bool for logical entities. We deal with the rest of the arithmetic types only when we have to."
Bjarne Stroustrup - Programming principles and practise using C++ - Chapitre 24 Numerics - p. 858
Ce qui inciterait à remplacer tes