A month ago when I first developed a non-commercial project using ReasonML, I got a new experience, is a functional programming. This language is an alternative syntax from OCaml. OCaml itself is a purely functional language, the features offered are interesting. For example: type inference, strongly type system, algebraic datatypes, and many more. Interesting right?
Now after trying that, I began to interest functional programming. Finally i tried to implement the functional paradigm in a different language, namely Rust.
Introduction
Functional programming (FP) is a programming paradigm which allows us to write expressive, concise, and elegant code. Functional programming also helps developers to manage code so that it doesn't DRY (Don't Repeat Yourself) that's mean doesn't write the same code over and over again. Other functional languages for example like Lisp, Clojure, Erlang, Haskell, R, and many more.
Okay, but why Rust?
The question is, is a Rust functional language? the answer is, no. Although Rust himself was inspired by ML family of language, Rust it's not functional. But fortunately Rust has several features that are similar to other functional languages, such as: algebraic datatypes, expressive types, and others.
Tables of Contents
Primitive Types
In order not to jump right away, it would be nice if we had to know several data types in Rust. This also applies to all programming languages.
Booleans
The most basic datatype is the simple true / false
value, in Rust it's called bool
let x = true;
let y: bool = false;
Char
The char
data type has a single Unicode value. We can use char
data type with a single tick ('
)
let x = 'x';
let two_hearts = '💕';
Unlike some other languages, char
in Rust is not one byte, but four.
Numeric Types
Rust has several numeric type category variants, such as signed(i
) and unsigned(u
), fixed size (8
, 16
, 32
, 64
) and variable(isize
, usize
) types.
let x = 42; // `x` has type `i32`.
let y = 1.0; // `y` has type `f64`.
Arrays
Like many other programming languages, Rust also has an array data type. By default, arrays in Rust can't be changed. Unless you initialize it with mut
let a = [1, 2, 3]; // a: [i32; 3]
let mut m = [1, 2, 3]; // m: [i32; 3]
Functions
Functions also have data types! For example like:
fn foo(x: i32) -> i32 { x }
let x: fn(i32) -> i32 = foo;
In this case, the foo ()
function has a return type numeric: i32
, and returns the valuex
.
For more information, you can check here: primitive types
Closures
Closure is a mechanism by which an inner function will have access to the variables defined in its outer function’s lexical scope even after the outer function has returned. Up to here understand? in short closures is an inner function that has access to retrieve a value throughout the scope both inside and outside.
fn fmt(prev_str: &str) -> String {
let mut new_str = String::new();
let closure_annotated = |next_str| -> String {
new_str.push_str(prev_str);
new_str.push_str(next_str);
return new_str;
};
closure_annotated("dolor sit amet")
}
let r_txt = "Lorem ipsum ";
assert_eq!("Lorem ipsum dolor sit amet", fmt(r_txt));
In this case, in new_str.push_str ()
section where closure_annotated
accesses thenew_str
variable then changes the value.
Currying
Currying is a process in functional programming in which we can transform a function with multiple arguments into a sequence of nesting functions. It returns a new function that expects the next argument inline.
#[derive(Debug)]
struct States<'a> {
a: &'a i32,
b: &'a i32,
}
trait Currying {
type ReturnType: Fn(i32) -> i32;
fn add(self) -> Self::ReturnType;
}
impl Currying for States<'static>{
type ReturnType = Box<dyn Fn(i32) -> i32>;
fn add(self) -> Self::ReturnType {
Box::new(move|x| {
x * self.a
})
}
}
let r_value: States = States {
a: &100,
b: &100
};
let r1 = r_value.add();
let r2 = r1(5);
assert_eq!(500, r2);
There are 2 parameters here, namely a
,b
where each has a numeric data type, then in trait
section is a function interface, a place for initializing functions. These traits are similar to typescript interfaces.
Higher Order Functions(HOF)
Higher order functions are functions that use other functions as parameters or as a result of returns.
fn map<F>(arr: &[i32], func: F) -> Vec<i32> where F: Fn(&i32) -> i32{
let mut new_array: Vec<i32> = vec![];
for i in arr.iter() {
new_array.push(func(i))
}
return new_array
}
let lists = vec![1, 4, 9, 16];
let result = map(&lists, |i| *i + 2);
assert_eq!(vec![3, 6, 11, 18], result)
So func
andmap
are higher order functions, where this function is used to change every contents of an array. The return result is a new array of the same length as the modified originalArray
.
Lazy Evaluation
Lazy evaluation or non-strict evaluation is a process of holding the evaluation of an expression/function
until the value is needed. The goal is to avoid repeated evaluations.
struct State {
x: i32,
}
trait Lazy {
fn add(&self) -> i32;
fn multiply(&self) -> i32;
fn add_or_multiply(&self, add: bool) -> i32;
}
impl Lazy for State {
fn add(&self) -> i32 {
println!("executing add");
&self.x + &self.x
}
fn multiply(&self) -> i32 {
println!("executing multiply");
&self.x * &self.x
}
fn add_or_multiply(&self, add: bool) -> i32 {
match add {
true => self.add(),
false => self.multiply(),
}
}
}
let val: State = State {
x: 20
};
assert_eq!(40, val.add_or_multiply(true));
assert_eq!(400, val.add_or_multiply(false));
Functional programming (FP) provides many advantages, and its popularity has been increasing as a result. However, each programming paradigm comes with its own unique jargon and FP is no exception. By providing a glossary, i hope to make learning FP easier✌️
Source code: rustfp.github
Top comments (9)
This is not the definition of currying in computer science.
Currying is (in maths and computer science) a programming technique that transforms a function of N parameters into a chain of functions (N functions) that each takes one parameter (a function of arity 1).
Whoa, yes i missconception about currying, thankyou for your comment, i will edit this ;)
Finally:
I was wondering how this code work as tail rec function and I discovered that it is not recursive at all.
You could tell by asserting like this:
I have been avoiding rust as too low level, but lately I have been wondering if it has a shot at becoming the first widely used backend FP ecosystem, where scala, clojure, and haskell have seemed to plateau and given up on widespread adoption.
An immediate question is "Why Functional programming in Rust?".
Rust's model is a different approach which solves many problems that FP solves. Accidental side effects don't normally happen in Rust and data are immutable by default. You're kind of forced to a modular and maintainable design too and risk-free concurrent execution comes easily.
So, why would you need or want FP in Rust ? Just as a kind of more elegant style ?
Omg, rust code looks so clean! Thanks for this article.
How is the last one lazy evaluation? The values are being evaluated eagerly in that example
I think OCaml is mostly pure but it could be impure in some cases. You can read here: ocaml.org/learn/tutorials/function...