The data type * real* (also * floating point* or
* float*)
is an approximation of a real number. Real numbers include numbers such as
* * and * 2 * that cannot be expressed as rationals, as well as
numbers such as ` 0.1` that cannot be represented exactly as binary
fractions. Since a real number is only approximated in the computer, such
numbers are called * inexact*.
Although floating-point numbers
are inexact, they can be very accurate, with about 15 decimal digits of
accuracy in a 64-bit floating point number.

Real constants are written with a decimal point and optionally with a
power-of-ten exponent in scientific notation: ` 1.0`, ` 0.3`,
` 9.11E-31` = * 9.11 10^-31*.

Operations on reals include:

`
`

` atan` can take one argument (in radians) or two:

` (atan y x)` computes * tan^-1( y / x) *.

Operations such as ` +` can be performed on combinations of exact
and inexact numbers, producing inexact results (``inexactness is
contagious'').