| realn | n-byte real number type |
| Syntax |
| ||||||||||||||||||||||||
| Description | The realn (n-byte real number) types are machine-dependent types that occupy a specified number of bytes. By contrast, the real type is, in principle, a machine-independent and mathematical type (however, it overflows when the exponent of the value is too large or small and it has only a limited amount of precision).
| ||||||||||||||||||||||||
| Example |
var width : real4
var height : real8
| ||||||||||||||||||||||||
| Details | Turing implements the type real using 8 byte floating point representation. This provides 14 to 16 decimal digits of precision and an exponent range of at least -38 .. 38. The PC and Macintosh versions of Turing have 16 decimal digits of accuracy because they use IEEE standard floating point representation. This implies that real8 and real are essentially the same type, so in practice there is no advantage to using real8 rather than real. However, real4 has the advantage of occupying half as much space (with correspondingly reduced precision). Arithmetic for all real types (real, real4 and real8) is carried out with the accuracy and exponent range of 8-byte reals. The type real4 is sometimes called single precision (because it occupies a single 4-byte word) and real8 is sometimes called double precision.
| ||||||||||||||||||||||||