[Tutor] Rounding a float to n significant digits

[wesley chun]

i think on a 32-bit platform, C doubles (IEEE754) are limited to 10 **
308.25 which is pretty close to 2 ** 1024

-wesley