Download as PDF, PPTX
More Related Content
![General OOP concept [by-Digvijay]](https://cdn.slidesharecdn.com/ss_thumbnails/cpresentationdvj-131231040948-phpapp01-thumbnail.jpg?width=600ounds&width=560&fit=bounds)
Similar to Introduction to functional programming (In Arabic)
Introduction to functional programming (In Arabic)
- 1.
- 2.
- 3. Hello! I am OmarAbdelhafith from Jordan I live in Dublin I write blogs on http://nsomar.com I work with computers @Zendesk I write mainly Swift, Haskell and Elixir On twitter @ifnottrue On github @oarrabi I organize Swift and Elixir meetups (in Dublin)
- 4.
- 5.
- 6. Intro to functionalprogramming • What topics to discuss? • lambda calculus • practical category theory • functors, monad, monoids • promises, futures, next • frp, reactive • fp langs
- 7. Functional programming First item: Introto functional programming
- 8. To start: Introto functional programming • History • Definition • Some Examples • PS: This is an interactive session
- 9. Note: Languages used •Java • Swift • Haskell • Pseudo language
- 10.
- 11. Functional programming Functional Programming(FP): Is a declarative programming paradigm, where programs are built around mathematical functions and immutable data transformation.
- 12. ^ That ismy definition
- 13. FP is avery wide field The definition can vary from lenient to rigorous
- 14. Lenient Definition Graham Hutton: Functionalprogramming is style of programming in which the basic method of computation is the application of functions to arguments; • No mention of immutability
- 15. In-the-middle Definition From Haskellwiki: In functional programming, programs are executed by evaluating expressions, ... Functional programming typically avoids using mutable state. ... (it) requires that functions are first-class, ...(to end of quote) • Mathematical functions is implied
- 16. Super Rigorus Definition Well,Not really a definition. But, Standarized ladder of functional programming (Lambda Conf)
- 19.
- 21. Lambda calculus • AlonzoChurch In 1930 invented λ-calculus • Calculation is done by lambda abstraction (definition) and binding (application) • There are no assignments of variables
- 25. • John McCarthycreated LISP in the 1950s • Inspired from lambda calculus • LISP is the first FP lang • Just a bit after Fortran • Introduced lots of concepts (macros, tree ds) • Still alive today (Common Lisp, Dialects, etc..)
- 27. • Created byan International comittee in 1987 • Pure functional • lazy • in 90s they implemented ideas from Category Theory • functors, monads, monoids
- 28.
- 30. Today • JVM: Scala,Clojure • Web: Clojure Script, ELM • BEAM VM: Erlang, Elixir • LLVM: Swift • General purpose: Idris
- 31. Recap? Is a declarativeprogramming paradigm, where programs are built around mathematical functions and immutable data transformation.
- 32. Decompose • Declarative programmingparadigm • Mathematical functions • Immutable data
- 33. Decompose • Declarative programmingparadigm • Mathematical functions • Immutable data • Everything is an expression • Functions are data
- 34. Programming paradigms • Imperative(C, C++, Java, Go): • state and control (if, for, while) • How to do • Declarative (Haskell, SQL) • logic but no control • What to do
- 35. Imperative Programming Paradigm • Tellthe computer how to do things • Change state • Based on statment
- 36. Imperative Programming Paradigm • Procedural(Go, C): • functions and data • Object Oriented Programming (Java, C++, C#): • objects with behaviour and state • Tell dont ask
- 37. Declarative Programming Paradigm • Tellthe computer what to do • High level • Less control
- 38. Declarative Programming Paradigm • Database query • Regex • DSLs • Functional programming
- 39.
- 40.
- 41.
- 43.
- 45.
- 47.
- 49.
- 52. Decompose • Declarative programmingparadigm ✅ • Mathematical functions • Immutable data • Everything is an expression • Functions are data
- 53. Mathematical functions • Theyare pure • They are Total • They are composable
- 55. Pure functions • Givensame input, return same output • They dont cause side effects
- 56. Pure functions: Another definition Functionsthat you can memoize (memoization) -- Bartosz Milewski
- 57. A bit offtopic- Memoization • Technique to speedup programms by storing results • Similar to caching but low level • Tabulate
- 58. Memoization Example: Fibonacci Withoutmemoization
- 59. • fib(n: 10)-> 88 times • fib(n: 11) -> 143 times
- 60. Memoization Example: Fibonacci Withmemoization
- 62. • fib(n: 10)-> 9 times • fib(n: 11) -> 10 times
- 64. Pure functions: functions thatcan be memoized
- 65. Pure functions • sin(x)❓ • length(s) ❓ • add(n1,n2) ❓ • random() ❓ • printf("Hello World") ❓
- 66. Pure functions • sin(x)✅ • length(s) ❓ • add(n1,n2) ❓ • random() ❓ • printf("Hello World") ❓
- 67. Pure functions • sin(x)✅ • length(s) ✅ • add(n1,n2) ❓ • random() ❓ • printf("Hello World") ❓
- 68. Pure functions • sin(x)✅ • length(s) ✅ • add(n1,n2) ✅ • random() ❓ • printf("Hello World") ❓
- 69. Pure functions • sin(x)✅ • length(s) ✅ • add(n1,n2) ✅ • random() ❌ • printf("Hello World") ❓
- 70. Pure functions • sin(x)✅ • length(s) ✅ • add(n1,n2) ✅ • random() ❌ but random(seed) ✅ • printf("Hello World") ❓
- 71. Pure functions • sin(x)✅ • length(s) ✅ • add(n1,n2) ✅ • random() ❌ but random(seed) ✅ • printf("Hello World") ❌
- 72. Mathematical functions • Theyare pure ✅ • They are Total • They are composable
- 73. Total functions • Forall set of inputs return always an output • add :: (Int, Int) -> Int
- 74. Examples • add(int, int),sub(int, int) ❓ • mul(int, int), div(int, int) ❓ • is_even(int) ❓ • int parseInt(String s) throws NumberFormatException ❓ • FILE *fopen(const char *filename, ...) ❓
- 75. Examples • add(int, int),sub(int, int) ✅ • mul(int, int), div(int, int) ❓ • int parseInt(String s) throws NumberFormatException ❓ • FILE *fopen(const char *filename, ...) ❓
- 76. Examples • add(int, int),sub(int, int) ✅ • mul(int, int) ✅, div(int, int) ❌ • int parseInt(String s) throws NumberFormatException ❓ • FILE *fopen(const char *filename, ...) ❓
- 77. Examples • add(int, int),sub(int, int) ✅ • mul(int, int) ✅, div(int, int) ❌ • int parseInt(String s) throws NumberFormatException ❌ • FILE *fopen(const char *filename, ...) ❓
- 78. Examples • add(int, int),sub(int, int) ✅ • mul(int, int) ✅, div(int, int) ❌ • int parseInt(String s) throws NumberFormatException ❌ • FILE *fopen(const char *filename, ...) ❌
- 79. Mathematical functions • Theyare pure ✅ • They are Total ✅ • They are composable
- 80. Function composition • f:: a -> b • g :: b -> c • g(f(a)) • f o g :: a -> c
- 81. Function composition • multiplyBy2:: Int -> Int • convertToString :: Int -> String • x = multiplyBy2(2) • convertToString(x) • convertToString(multiplyBy2(2)) • (convertToString . multiplyBy2)(2)
- 82. Example in FPlanguage Let's delay it after high order functions
- 83. Decompose • Declarative programmingparadigm ✅ • Mathematical functions ✅ • Immutable data • Everything is an expression • Functions are data
- 84. Immutable data • Datadoes not change • it gets copied when passed • In FP, all data is immutable
- 87. What about performance No problemwith "Copy on write"
- 88. Decompose • Declarative programmingparadigm ✅ • Mathematical functions ✅ • Immutable data ✅ • Everything is an expression • Functions are data
- 89. Everything is anexpression • Statments: tell how to do • Expression: return things
- 90.
- 91.
- 92. Everything is anexpression • Statments: tell how to do • Expression: return things • In FP: There is no void • In Haskell ---> absurd :: Void -> a
- 93. Decompose • Declarative programmingparadigm ✅ • Mathematical functions ✅ • Immutable data ✅ • Everything is an expression ✅ • Functions are data
- 94. Functions are data •In FP functions are data • Higher-order function • They can be passed to functions • They can be returned from functions
- 95. They can bepassed as params
- 99. They can bereturned
- 103.
- 106. Decompose • Declarative programmingparadigm ✅ • Mathematical functions ✅ • Immutable data ✅ • Everything is an expression ✅ • Functions are data ✅
- 107. Recap • Declarative: Whatto do and not how to do it • Mathematical functions: Pure, Total, Composable • Immutable data: Dont change the state • Everything is an expression: and they are composable • Functions are data: Can be passed and returned from functions
- 108. Why are weseeing a rise in FP • Multi core and paralellism • Determinism, testability • They align well with cloud computing and microservices
- 109.
- 110.