summaryrefslogtreecommitdiffstats
path: root/static/talks/fp-js/index.org
blob: 8a4bf6ce3a29319917dbb79ddbea545b00a60e44 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
#+TITLE: Functional Programming in JavaScript
#+PROPERTY: :html-toplevel-hlevel 1
#+PROPERTY: :with-toc 0
* Why?

Imperative programming is concerned with *how*

Functional programming is concerned with *what*

* Concepts

 - First-class Functions
 - Higher-order Functions
 - Recursion
 - Pure functions
 - Currying & Partial Application

* Further concepts

 - Lazy Evaluation
 - Types & Data Structures
 - Category Theory

* First-class functions

 - Have no restriction on their use
 - Are values
 - Enable the use of callback functions in JavaScript



#+BEGIN_SRC js
const fn = function () {
  return 2
}
#+END_SRC


* Higher-order functions

Functions that operate on other functions are higher-order functions



#+BEGIN_SRC js
const succ = function (x) {
  return x + 1
}

const arr = [1, 2, 3, 4]

arr.map(succ)
#+END_SRC

Here, =Array.prototype.map= is the higher-order function

* Higher-order functions (cont.)

Functions that return functions are also higher-order functions



#+BEGIN_SRC js
function adder (n) {
  return function (x) {
    return n + x
  }
}

const add1 = adder(1)
#+END_SRC

=adder= is a higher-order function

* Pure functions

Functions without side-effects

#+BEGIN_SRC js
const succ = (x) => x + 1

console.log(succ(succ(1)))

// could be optimised away by a compiler, e.g.:

console.log(3)
#+END_SRC

* Recursion

Functions that call themselves

#+BEGIN_SRC js
function fibonacci (n) {
  switch (n) {
    case 0:
    case 1:
      return 1
    default: 
      return fibonacci(n - 1) + fibonacci(n - 2)
  }
}
#+END_SRC

* Partial application

The infamous =Function.prototype.bind= in JavaScript

#+BEGIN_SRC js
function add (x, y) {
  return x + y
}

const add1 = add.bind(add, 1)

add1(3) // = 4
#+END_SRC

* Partial application (cont.)

After ES6 introduced arrow functions, partial application has become
more popular

#+BEGIN_SRC js
const add = x => y => x + y
#+END_SRC

* Currying

Related to partial application, but more implicit and general

Translates */1/ function of arity /n/* to */n/ functions of arity /1/*

#+BEGIN_SRC js
function volume (w, d, h) {
  return w * d * h
}

const vol = curry(volume)
vol(10)(20)(30)
// is strictly equivalent to
volume(10, 20, 30)
#+END_SRC

* Easy Currying

In order to make currying (and partial application) easier to use,
move the *most important* argument to a function to the end:

#+BEGIN_SRC js
const badMap = (arr, fn) => arr.map(fn)
const goodMap = (fn, arr) => arr.map(fn)
const curriedBadMap = curry(badmap)
const curriedGoodMap = curry(goodMap)

const goodDoubleArray = goodMap(x => x * 2)
const badDoubleArray = badMap(_, x => x * 2)
#+END_SRC

The bad version requires the curry function to support a magic
placeholder argument and doesn't look as clean.

* Practical Currying

Currying is not automatic in JavaScript, as in other languages

External tools don't (currently) to statically analyse curried
functions

Solution: Don't expose curried functions
Instead, write functions as if currying were automatic

* Functional composition

Creating functions from other functions

Usually provided by =compose= (right-to-left) and =pipe= (left-to-right)

A very simple definition of =compose= for only two functions would look like this

#+BEGIN_SRC js
function compose (f, g) {
  return function (...args) {
    return f(g(...args))
  }
}
#+END_SRC

* Functional composition (cont.)

#+BEGIN_SRC js
const plusOne = x => x + 1
const timesTwo = x => x * 2

const plusOneTimesTwo = compose(timesTwo, plusOne)
const timesTwoPlusOne = compose(plusOne, timesTwo)

nextDoubled(3) // = (3 + 1) * 2 = 8
timesTwoPlusOne(3) // = (3 * 2) + 1 = 7
#+END_SRC

* pipe

What about =pipe=?

=pipe= does the same thing, but runs the functions the other way around

=pipe(f, g)= is the same as =compose(g, f)=

* Point-free programming

With currying and higher-order functions, we (often) don't need to declare function arguments

#+BEGIN_SRC js
const modulo = a => b => b % a
const eq = a => b => a === b

const isEven = x => eq(0)(modulo(2)(x))
const isEvenPointFree = compose(eq(0), modulo(2))
#+END_SRC

* Further Resources

- [[https://drboolean.gitbooks.io/mostly-adequate-guide/content/][Mostly adequate guide to FP (in javascript)]]
- [[http://ramdajs.com/][Ramda]], a general-purpose FP library
- [[https://sanctuary.js.org/][Sanctuary]], a JavaScript library for Haskellers