Page: 10 
Example 
The number 2.5 is represented exactly in double precision IEEE
floating point notation as
since it can be written as 
The number 2.4 is represented as
However, it cannot be represented exactly because it can't be written as a sum of powers of two in a range allowed by the fixed word length. 
The error caused by rounding 2.4 to its floatingpoint representation is
less than 9x10^{17}. The exact decimal representation of the
binary floatingpoint value is
