4.8 Floating-Point Numbers

Real numbers are numbers that have a fractional part. Because of the way they are stored internally, real numbers are also known as floating-point numbers. The numbers 5.5, 8.3, and -12.6 are all floating-point numbers. C++ uses the decimal point to distinguish between floating-point numbers and integers, so a number such as 5.0 is a floating-point number while 5 is an integer. Floating-point numbers must contain a decimal point. Numbers such as 3.14159, 0.5, 1.0, and 8.88 are floating-point numbers.

Although it is possible to omit digits before the decimal point and specify a number as .5 instead of 0.5, the extra 0 makes it clear that you are using a floating-point number. A similar rule applies to 12. versus 12.0. Floating-point zero should be written as 0.0.

Additionally, a floating-point number may include an exponent specification of the form e±exp. For example, 1.2e34 is shorthand for 1.2x10³⁴.

The form of a floating-point declaration is:

float variable;   // comment

Again, there is a limit on the range of floating-point numbers the computer can handle. The range varies widely from computer to computer. Floating-point accuracy is discussed further in Chapter 19.

Floating-point numbers may be output using std::cout. For example:

std::cout << "The answer is " << (1.0 / 3.0) << "\n";