Do you have any examples where writing the expression for the number may lead to floating point erros, while using scientific notation does not?