Can someone give me an example of a floating point number (double precision), that needs more than 16 significant decimal digits to represent it?
I have found in this thread that sometimes you need up to 17 digits, but I am not able to find an example of such a number (16 seems enough to me).
Can somebody clarify this?
