A hexadecimal floating-point literal may not begin with any of the alphabetic hexadecimal characters or it will be interpreted as an identifier; such numbers must be prefixed with the 0 digit.
The literal must specify the correct number of hexadecimal digits according to the size of the real-number data-type to which it will be assigned. For REAL4, REAL8, and REAL10 variables, the respective number of digits in the literal must be 8, 16, and 20. For literals encoded with a leading zero, the respective number of digits must be 9, 17, and 21.