To store/parse floats the single precision floating point format is used.
http://en.wikipedia.org/wiki/Single_precision