Encoding: Real numbers (floating point)

1. 0.18 = 18.10^(-2) = two integers (18,-2) called mantissa and exponent

• float = 4 octets: 1 sign bit, 23 bits for mantissa, 8 bits for exponent
• double precision = 8 octets: 1 sign bit, 52 bits for mantissa, 11 bits for exponent

2. Same operators than int plus specific functions found in math.h library (described latter): sin, cos, exp...

Eg. man sqrt

11 - 15