What functionalities are you thinking about? Proper float handling to prevent the '0.1 + 0.2 = 0.30000000004' bug? Manual computation of square roots/sinus functions?