Plange is an open-source development suite. The language consists of semantics from imperative, functional, and constraint-based programming. ODE solving, first-order logic, and type checking are examples of the available algebra and calculuses. A hybrid memory model supports garbage collection and native pointers. Many abstractions are available including inline multi-platform assembler, algebraic types, database records, code-level reflection and docstrings. The IDE communicates the success of verification, compilation, and validation steps to the user though real-time feedback, and also displays the results of symbolic workloads. This web domain hosts documentation, standard and third-party import libraries, cloud resources, a support community, and more.

print("Hello, world!");

End of line comment (red is comment text)

print("My name is HAL 9000."); //only kidding!

Inline comment (red is comment text)

getRandomNumber := { return 4; /*choosen by fair dice roll*/ }; //guaranteed to be random

print the voltage across a discharging capacitor

V := coerce<Real>(input("Volts: "));
R := coerce<Real>(input("Ohms: "));
C := coerce<Real>(input("Farads: "));
𝑑 := coerce<Real>(input("Seconds: "));

𝜏 := R * C;
print(V * 𝑒^(-𝑑/𝜏));

Variables can be reassigned.

Reassign a variable

color ← "Blue";
color ← "Red";

Type constraint on a variable

<Number> x ← 1337;

Memory Model

x ← 2.718281;
<Pointer<Float>> x_ptr = address_of(x);

Symbolic manipulation

x = 1337;
tan(y*2) = x; // y is a free variable
echo(y); // arctan(1337) / 2 = { 1.570048, -1.571544 };


A symbol with an immutable value is a constant.



Ο€ or pi is a predefined constant, and provides arbitrarily high precision in symbolic manipulation.

Constants are created using the definition operator :=


daysInAWeek := 7;


Function are created with parenthesis ( ) containing the parameter list, and curly braces { } containing the implementation. Nullary functions may elide their parameter list.


doubler := (x) { return x * 2; };

Functions have arity.

A binary function

geometric_mean := (x, y) { return √(x * y); };


Constants and variables may be strongly typed implicitly or explicitly, late bound, or duck typed.

Explicit strong typing

<Int> x ← 10;

Int or β„€ is shorthand for integer. The first line constrains x to Int values. See Type System.

Functions Types

Create function types using the β†’ operator, or ->.


<Int β†’ Int> doubler = (x) { return x * 2; };

doubler will only accept an Int as input, and will only return an Int. Function arguments may also be given types.

Semantically equivalent to the previous example

doubler := (<Int> x) { return x * 2; };

Variables that have no specified type constraint are dynamically typed.

Assigning objects of varied type to a variable

x ← 10;
x ← "Alice";
x ← { print("fubar"); };

Making Types

The type keyword (not capitalized) is used to make a new Type object (capitalized).


Color := type {
	<Double> r;
	<Double> g;
	<Double> b;

<Color> red ← (| 1, 0, 0 |);

print(type_of( (| 1, 0, 0 |) ));  // output: Tuple<Number, Number, Number>
print(type_of(red));          // output: Color
print(type_of(Color));        // output: Type

Parametric Types

Parametric types are functions that return Type objects.


Node := (<Type> valueType) {
	return type {
		<valueType> v;
		Maybe<Node<valueType>> next;

Constant folding evaluates some invocations of type functions at compile time. Functions that return Type objects (or another type function) should be called with the angle bracket syntax:

Example invoking List using angle bracket syntax

<List<Int>> myList;

Pattern Matching

Pattern matching decomposes values into unbound symbols. Patterns are tested sequentially in the order given.

Tail recursive function to print the last element of a list

<List<_> β†’ Void> printLast := 
	(_ & tail) { printLast(tail); } |
	(x) { print(x); };

myList := [ 5, 12, 8, 9 ];



The prepend operator & takes a value on the left, and a list on the right. In the examples above the first parameter to the function is being broken apart into two pieces.

Note the use of the underscore _ character. It's substituted for a variable when the code doesn't care about the value. In the first line of the example above, we are unconcerned with the type of the elements the input list contains, and only need to ensure that the input is a list of something. In the second line, we don't need to know the value of the head element. The underscore keyword is called dont_care.

Another recursive function to filter values

filter := 
	(head & tail, predicate) {
		return predicate(head) ?
			head & this_func(tail, predicate) :
			this_func(tail, predicate);
	} |
	([], predicate) { return []; };


The inheriting keyword, used in conjunction with the type keyword, makes a new Type object inheriting the members of the specified base Types. See also: implementing


// base Type
Widget := type {
	<Void β†’ Image> Paint;

// derived Type
TextBox := type inheriting Widget {
	<String> text ← "Hello, world!";

	// override the inherited Paint method
	Paint ← (<Canvas> c) {
		return c.print(text);

Algebraic Types

Types can be combined together to make algebraic Types using the compound operator |.


Some := (t) { return type { <t> value; }; };
None := type {};
Maybe := (t) { return Some<t> | None };

<Void β†’ Maybe<Int>> get_age := {
	return coerce(input("What's your age? You don't have to tell me."));


Constraint solving

Many interesting problems may be constructed as one or more constraints using operators and functions. When an appropriate normal form or canonical form is given, constant folding, satisfiability solving, and symbolic manipulation yield equality constraints upon termination. See the computer algebra systems page. All the results in this section have been computed manually for demonstrative purposes.


children := {| abe, dan, mary, sue |};
ages := {| 3, 5, 6, 9 |};
children ↔ ages; // One child per one age (bijection operator)

abe > dan; //abe is older than dan
sue < mary; //sue is younger than mary
sue = dan + 3; //sue's age is dan's age plus 3 years
mary > abe; //mary is older than abe


The output when ran:



Example of a linear system

<Real * Real * Real β†’ Real> linear_interpolation :=
	(min, max, x) { min * (1 - x) + max * x }
inverted_linear_interpolation := (min, max, interpolated) { 
	interpolated = linear_interpolation(min, max, x); // solve for free variable x
	return x;

linear_map := (minIn, maxIn, v, minOut, maxOut) {
	v = linear_interpolation(minIn, maxIn, x); // solve for free variable x
	return linear_interpolation(minOut, maxOut, x);

assert(x = inverted_linear_interpolation(y, z, linear_interpolation(y, z, x));

One well studied domain is initial value problems. An ordinary differential equation is given with boundary conditions on free variables:

Example: Aerodynamic Drag On A Projectile

projectilePosition := (
		<Vector3D> initialPos,
		<Vector3D> initialVel,
		<Real> mass,
		<Real> drag,
		<Vector3D> gravity,
		<Real> t
) {
	// declare the position function, x
	<Real β†’ Vector3> x;
	// model x as a differential equation
	mass * Ξ”^2x(t)/Ξ”t^2 = -drag * Ξ”x(t)/Ξ”t + mass * gravity;
	// with boundary conditions
	x(0) = initialPos;
	Ξ”x(0)/Ξ”t = initialVel;
	// solve, substitute, evaluate
	return x(delta_t);

ODE solving gives a symbolic solution for x such that the following program is functionally equivalent. This constant folding is performed and cached at compile time.

Example (continued)

projectilePosition := (
		<Vector3D> initialPos,
		<Vector3D> initialVel,
		<Real> mass,
		<Real> drag,
		<Vector3D> gravity,
		<Real> t
) {
	a := 𝑒^(drag*t/mass);
	return (
			gravity * (mass-(mass*a + drag*t)) + 
			initialPos*a*drag^2 + 
		) / (a*drag^2);


Constraint solving is intractible in the general case. Users must familiarize themselves with the capabilities of the language, which are expected to expand. A demonstration of a semantically correct but nonfunctional program is in order.

Counter Example

<Collection * BinaryRelation β†’ Collection> sort := (items, ordering) {
	result ↔ items; // result and items make a bijection
	βˆ€ { ordering(result[i - 1], result[i]) | i ∈ (0...|result|) }; //the result has to be sorted
	return result; // solve, substitute, and return

The above function, sort, is functionaly equivalent to the sorting functions. However, this constraint based problem is not yet solvable using available techniques.

Further reading

This website does not use cookies or javascript, excluding the Google search bar and Google analytics.
copyright Β© Brent Lewis 2013-2018