Floating Point

The essential idea of floating point representation is that a fixed number of bits are used (usually 32 or 64) and that the binary point "floats" to where it is needed in those bits. Of course, the computer only holds bit patterns. Some of the bits of a floating point representation must be used to say where the binary point lies.

Floating point expressions can represent numbers that are very small and numbers that are very large. When a floating point calculation is performed, the binary point floats to the correct position in the result. The programmer does not need to explicitly keep track of it.