Skip to main content
Version: 1.7.0

Numbers

In LIGO, there are two types of numbers: integers and natural numbers.

  • Integer literals are the same found in mainstream programming languages, for example, 10, -6 and 0, but there is only one canonical zero: 0 (so, for instance, -0 and 00 are invalid).

  • Natural numbers are written as digits followed by the suffix n, like so: 12n, 0n, and the same restriction on zero as integers applies: 0n is the only way to specify the natural zero.

Contrary to integral numbers in other programming languages, numbers in LIGO have arbitrary-precision, that is, they do not overflow or underflow.

Digits of large numbers can be separated by an underscore, to increase readability.

// The following are integers
const zero = 0
const million = 1_000_000 // Grouping in French
const baekman = 100_0000 // Grouping in Korean
// The following are natural numbers
const zero_nat = 0n
const million_nat = 1_000_000n
const baekman_nat = 100_0000n

As a form of documentation, a type can be ascribed to each constant:

const zero : int = 0
const million : int = 1_000_000
const baekman : int = 100_0000
const zero_nat : nat = 0n
const million_nat : nat = 1_000_000n
const baekman_nat : nat = 100_0000n

Casting

In mathematics, natural numbers are a strict subset of integers, and can be used in any context where an integer is expected. In LIGO, this property does not hold true in general. Instead, a given binary arithmetic operation, say, is defined four times, so it can apply to any combination of natural numbers and integers: this is called overloading, and some programming languages extend it to user-defined functions (e.g. members in C++) -- but not LIGO.

So there are no implicit type casts in LIGO, but we can explicitly cast natural numbers to integers (this is safe in all contexts where an integer is valid) by calling the predefined function int. The inverse cast, from int to nat is called in mathematics the absolute value, or abs in LIGO.

const one : int = int(1n); // Explicit cast from nat to int
const two : nat = abs(2); // Explicit cast from int to nat

Adding

Addition in LIGO is accomplished by means of the + binary operator, which is overloaded to apply to any combination of natural numbers and integers, as shown in the following examples. Note that adding an integer to a natural number produces an integer, because the compiler cannot determine, in general, whether the result would be always a natural number for all inputs.

const a : int = 5 + 10; // int + int yields int
const b : nat = 5n + 10n; // nat + nat yields nat
const c : int = 5n + 10; // nat + int yields int
const d : int = 10 + 5n; // int + nat yields int
// const error : nat = 5n + 10;

Subtracting

Subtraction in LIGO is accomplished by means of the - binary operator which is overloaded to apply to any combination of natural numbers and integers, as shown in the following examples. The rule when subtracting two natural numbers is that the result is an integer because, in general, the compiler cannot determine whether the value of an expression is positive or zero for all inputs.

const a : int = 5 - 10; // int - int yields int
const b : int = 5n - 2n; // nat - nat yields int
const c : int = 10n - 5; // nat - int yields int
const d : int = 5 - 10n; // int - nat yields int
// const error : nat = 5n - 2n;

Negating

The arithmetic negation of a number is the same as subtracting that number from zero, so the negation of a natural numbers yields an integer:

const a : int = -5; // - int yields int
const b : int = -5n; // - nat yields int
// const error : nat = -5n;

Multiplying

Multiplication in LIGO is accomplished by means of the * binary operator which is overloaded to apply to any combination of natural numbers and integers, as shown in the following examples. The type rules for multiplication are the same as for the addition:

const a : int = 5 * 10; // int * int yields int
const b : nat = 5n * 2n; // nat * nat yields nat
const c : int = 10n * 5; // nat * int yields int
const d : int = 5 * 10n; // int * nat yields int

Dividing

Because LIGO features neither floating-point nor fixed-point arithmetic, division in LIGO is Euclidean. The predefined binary operator / returns the quotient and is overloaded like the multiplication. Of course, division by zero triggers an exception that interrups the execution, so the programmer must make sure this case cannot happen because the compiler cannot determine, in general, if a variable will have a given value (or not) for all inputs.

const a : int = 10 / 3; // int / int yields int
const b : nat = 10n / 3n; // nat / nat yields nat
const c : int = 10n / 3; // nat / int yields int
const d : int = 10 / 3n; // int / nat yields int

The binary operator % returns the positive modulo of the Euclidean division, that is, the following holds:

(n*(a/n)+(a%n) == a) && (0n <= a % n) && (a % n < abs(n))

It is overloaded as the Euclidean division / to allow for all four combinations of natural numbers and integers.

const a : nat = 120 % 9; // int % int yields nat
const b : nat = 120n % 9; // nat % int yields nat
const c : nat = 120n % 9n; // nat % nat yields nat
const d : nat = 120 % 9n; // int % nat yields nat

It is possible to obtain both the quotient and remainder together, by means of the predefined function ediv: See Euclidean division.