Functional programming techniques in regular JavaScript
1 like731 views
How to apply some functional programming ideas in regular JavaScript. 1. Immutability 2. Simple functions 3. Not that simple functions 4. Combining OOP and FP 5. Functors 6. Monads
![[FT-11][ltchen] A Tale of Two Monads](https://cdn.slidesharecdn.com/ss_thumbnails/monads-140308101321-phpapp01-140421092116-phpapp01-thumbnail.jpg?width=560&fit=bounds)
More Related Content
What's hot (10)
Viewers also liked (20)
Similar to Functional programming techniques in regular JavaScript (20)
![Multithreading and Parallelism on iOS [MobOS 2013]](https://cdn.slidesharecdn.com/ss_thumbnails/mobos2013-131122085438-phpapp02-thumbnail.jpg?width=560&fit=bounds)
More from Pavel Klimiankou (7)
Recently uploaded (20)
Functional programming techniques in regular JavaScript
- 2. ABOUT ME Pavel Klimenkov Senior Software Developer @ Geotab Inc. Oakville, ON, Canada
- 3. WHAT THIS TALK IS ABOUT • Pick few ideas from functional programming • Simplify as much as possible • Apply in JavaScript • PROFIT!
- 4. WHAT THIS TALK IS ABOUT • Pick few ideas from functional programming • Simplify as much as possible • Apply in JavaScript • PROFIT!
- 5. WHAT THIS TALK IS ABOUT • Pick few ideas from functional programming • Simplify as much as possible • Apply in JavaScript • PROFIT!
- 6. WHAT THIS TALK IS ABOUT • Pick few ideas from functional programming • Simplify as much as possible • Apply in JavaScript • PROFIT!
- 7. WHAT THIS TALK ISN’T ABOUT • OOP must die, FP will save the world • Detailed FP manual • “A monad is just a monoid in the category of endofunctors”
- 8. WHAT THIS TALK ISN’T ABOUT • OOP must die, FP will save the world • Detailed FP manual • “A monad is just a monoid in the category of endofunctors”
- 9. WHAT THIS TALK ISN’T ABOUT • OOP must die, FP will save the world • Detailed FP manual • “A monad is just a monoid in the category of endofunctors”
- 10. AGENDA •Immutability •Simple functions •Not that simple functions •Combining OOP and FP •Functors •Monads
- 11. WHAT IS FP? It’s a declarative programming style where all functions have no side effects and data is readonly. BTW, SQL is also declarative
- 12. 1. IMMUTABILITY Objects don’t change their state Values are constants
- 13. IMMUTABILITY • Use only read-only data • If something needs to be changed - create new instance instead
- 14. EXAMPLE Convert GPS coordinates to screen points
- 23. HOW TO REDUCE TEMPTATION • Use const or let instead of var • filter/map/reduce instead of for • Object.freeze - to prevent object mutation
- 24. IMMUTABILITY: PROS • Simpler code • Less bugs • Atomic object construction (object is either fully initialised, or doesn’t exist) • Less temporal coupling (when order of initialisation matters) • Simpler caching • Thread safety*
- 25. IMMUTABILITY: CONS • Higher CPU impact (more things to do) • Higher memory impact (and more GC) • Higher brain impact (OOP often comes with state)
- 26. 2. SIMPLE FUNCTIONS First class functions Pure functions
- 27. FIRST CLASS FUNCTIONS Can be passed as arguments Can be assigned as values
- 28. PURE FUNCTIONS • No side-effects • Depend on arguments only (deterministic)
- 32. SIMPLE FUNCTIONS: PROS • No side-effects • Less bugs • Simpler testing • Simpler code • Absolute thread safety
- 33. SIMPLE FUNCTIONS: CONS • No side-effects • Our job is to create side-effects • IO and Cache are side-effects
- 34. 3. NOT THAT SIMPLE FUNCTIONS Functions composition Functions currying
- 35. FUNCTIONS COMPOSITION Composition - gluing functions together
- 38. EXAMPLE: COMPOSITION Two temporary arrays. Oh, the horror!
- 39. EXAMPLE: COMPOSITION One temporary array. Harmony restored.
- 42. FUNCTIONS CURRYING Currying - partial application of a function.
- 49. NOT THAT SIMPLE FUNCTIONS: PROS • Flexibility in using and reusing functions • Functions tend to be smaller • Memoization (“Simple functions” PROS list was already too long)
- 50. NOT THAT SIMPLE FUNCTIONS: CONS • Sometimes code looks cryptic • Might introduce negative performance impact
- 51. COMBINING OOP AND FP Shell - OOP, IO, side-effects Core - FP, logic, pure functions Core Shell
- 52. MAIN PRINCIPLES • Imperative shell • Deals with IO • Thin layer
- 53. MAIN PRINCIPLES • Functional core • deals with immutable data and pure functions • contains application logic
- 55. EXAMPLE: MOVING VEHICLES IO, shell
- 56. EXAMPLE: MOVING VEHICLES Logic, pure functions, core
- 57. SHELL/CORE: PROS • Main bugs will come from the shell, which is small • Core is easy to test. Unit tests should be enough • Shell is the one that needs integration tests (shell is small!) • Command query responsibility segregation out of the box
- 59. FUNCTORS Smart containers for values
- 60. EXAMPLE: INCREMENTING A PICTURE
- 61. EXAMPLE: INCREMENTING A PICTURE
- 62. EXAMPLE: INCREMENTING A PICTURE
- 63. EXAMPLE: INCREMENTING A PICTURE
- 64. EXAMPLE: INCREMENTING A PICTURE ??? 3.png ???
- 65. FUNCTOR • Is a container around some data, • that knows how to apply functions to that data, • keeping result in the container
- 66. FUNCTOR • Container - array, async, nullable - anything • Some data - numbers, objects, other functors - anything • Functions are applied through map()
- 67. ARRAY IS A FUNCTOR!
- 68. PROMISE IS ALMOST A FUNCTOR!
- 70. USEFUL FUNCTORS • Maybe • Either • IO • Writer • State
- 71. MAYBE - ALMOST LIKE NULLABLE
- 72. MAYBE - ALMOST LIKE NULLABLE*
- 73. MAYBE - ALMOST LIKE NULLABLE
- 74. MAYBE EXAMPLE
- 75. MAYBE EXAMPLE
- 76. EITHER = LEFT ♥ RIGHT
- 77. EITHER = LEFT ♥ RIGHT
- 78. EITHER = LEFT ♥ RIGHT
- 79. EITHER = LEFT ♥ RIGHT
- 80. EITHER = LEFT ♥ RIGHT
- 81. EITHER = LEFT ♥ RIGHT
- 82. EITHER ♥ RAILWAY ORIENTED PROGRAMMING Right Left f(x) g(x) h(x) Picture and concept are shamelessly borrowed from here
- 83. TO BE HONEST… Traditional Maybe = Just ♥ Nothing
- 84. OTHER FUNCTORS • IO - for… IO • Writer - for logs • State - for… state
- 85. FUNCTORS: PROS • Convenient control over computation • Logic and context separation
- 86. FUNCTORS: CONS • Steep learning curve • Temptation to use functors for everything
- 87. SKIPPED CHAPTERS • Applicatives • Monoids
- 88. MONADS
- 89. MONADS ARE CURSED: EITHER YOU DON’T GET THEM, OR YOU CANT EXPLAIN THEM
- 93. MAYBE WILL BECOME A MONAD IF
- 94. MAYBE WILL BECOME A MONAD IF
- 95. MAYBE WILL BECOME A MONAD IF
- 96. NOW YOU KNOW KUNG FU MONADS Whoa!
- 97. ARRAY WILL BECOME A MONAD IF ECMAScript gets flatten or flatMap for arrays (soon)
- 98. BTW, PROMISE IS ALMOST A MONAD
- 100. SUMMARY • Immutability - everything is read-only, less bugs • Pure functions - no side-effects, less bugs • Composition, currying - reusing functions • Functors - separating context and logic • Monads - smart functors