Some remarks about the collatz problem