All the world's a float

I know the title of this post makes no sense, but, well, there it is... but seriously...

Over at *Coder Blog, Mike points out an excellent article at the ridiculous fish blog about the formatting of floating point numbers in memory. Didn't you know it wasn't particularly trivial to store fractions of a number with just ones and zeroes? I haven't thought much about the details for a few years, but here is a pretty thorough description, good demonstration.