Numbers in CIGAL can be either integers or floating point numbers. If a number contains a decimal point then it is stored as a 32-bit floating point number with approximately 7 significant digits. Floating point numbers are always decimal (base 10). Some examples of floating point numbers are:
10. 123.456 .0032 0.12

If a number is entered as an integer, without a decimal point, then it can be in base 2, 4, 8, 10, or 16. The base is determined by the presence of a letter at the end of the number. If there is no letter, then the base is determined by the value of the internal variable IBASE, which is 10 (decimal) by default (see IBASE[4]). However, you can specify the base explicitly for any number by ending the number with the letter 'b', 'q', 'o', 'd', or 'h' which specify base 2, 4, 8, 10, or 16 respectively. The following table gives some examples for each base with the corresponding decimal values in parentheses (the examples assume IBASE = 10):

Base Range of digits Examples Decimal values
==== =============== ======== ==============
2 0-1 10b, 101B, 100110b 2, 5, 38
4 0-3 10q, 31Q, 1231q 4, 13, 109
8 0-7 10o, 27O, 1321o 8, 23, 721
10 0-9 10d, 33, 3214 10, 33, 3214
16 0-f 10h, 0fH, 0b3ah 16, 15, 2874

Anytime that you enter a number in CIGAL you can enter the number in any base by using the base specifying characters. This is true whether you are entering numbers on a command line or as part of data input (see DATAENTRY[1]). You can freely intermix different bases, integers, or floating point integers from one number to the next.

Note: As usual for hexadecimal numbers, the single character values corresponding to 10-15 are expressed as the letters a-f. Because the command line parser identifies numbers by the fact that the first character is a digit, hexadecimal numbers that start with an a-f must be preceded by an initial '0' to avoid being interpreted as character strings. Letters that are part of a number can be either upper or lower case.

Numbers are stored within CIGAL in one of seven different formats. The storage format is determined by the type of variable (see VARIABLES[1]). The formats are:
Name Storage Range of legal values
==== ======= =====================
Bit unsigned 1 bit integer 0 - 1
Crumb unsigned 2 bit integer 0 - 3
Nibble unsigned 4 bit integer 0 - 15
Byte unsigned 8 bit integer 0 - 255
Integer signed 16 bit integer -32767 - 32767
Long signed 32 bit integer -2147483647 - 2147483647
Real signed 32 bit floating pt. 7 significant digits

The range of legal values for signed integers is one number less than the maximum range possible because the most negative value (-0: the sign bit set to 1, followed by all zeros) is reserved for a 'missing value' code (see ARITHMETIC[1]).

NOTE: CIGAL currently has some bugs when dealing with very large LONG integers; it tends to round them off to about 7 significant digits.