![Sum types (Scala)
sealed trait Maybe[+A]
case class Just[A](a: A)
extends Maybe[A]
case object None
extends Maybe[Nothing]](https://image.slidesharecdn.com/goodfunctionalprogramming-170412105926/85/Good-functional-programming-is-good-programming-54-320.jpg)
Good functional programming is good programming
- 1. (= (good (functional programming)) (good programming)) @KenScambler
- 3. λx → x+1
- 4. What is a functional programming language?
- 5. (lambda (x) (+ x 1))
- 6. do |x| x + 1 end
- 7. new Function<Integer,Integer>() { @Override public Integer apply(Integer x) { return x + 1; } };
- 8. You can do FP in any language!
- 9. John McCarthy 1927 - 2011 Lisp, 1958 (DEFUN FAC (N) (IF (= N 0) 1 (* N (FAC (- N 1))))) • Value-oriented • Garbage collection • Lambda functions
- 10. Guy Steele 1954- Scheme, 1970 (define (fac n) (if (= n 0) 1 (* n (fac (- n 1)))))) • Lexical scope • Tail call optimization • Continuations Gerald Sussman 1947-
- 11. Robin Milner, et al 1934 - 2010 ML, 1973 fun fac (0 : int) : int = 1 | fac (n : int) : int = n * fac (n - 1) • Statically typed • HM type inference • ADTs, Pattern matching
- 12. Haskell Committee Haskell, 1990 fac :: (Integral a) => a -> a fac n = if n < 2 then 1 else n * fac (n-1) • Purity • Lazy evaluation • Monads
- 13. Why do I care?
- 14. The things that make FP good are the same things that make good software good
- 15. Simple Modular Local reasoning Extensible Testable
- 16. Think better thoughts Your design ideas are hi-res
- 17. Think better thoughts Your ideas expressed in a language are low-res
- 18. If you only think “in a language”, then your ideas are already low res
- 19. Functions & composition Immutability Purity & types The big ideas λ Laziness
- 20. Functionsλ
- 23. Composition
- 24. A B
- 25. A B B C
- 26. A B B C C D
- 27. A C C D
- 28. A D
- 29. Equational reasoning 5 + (3 * 4)
- 30. Equational reasoning 5 + 12
- 33. Equational reasoning val x = expensiveThing() process(x,x,x)
- 34. Referential Transparency “An expression is said to be referentially transparent if it can be replaced with its corresponding value without changing the program's behavior.” - Wikipedia
- 35. It’s just good sense! Being able to reason about software is just good programming!
- 36. Immutability
- 37. case class Point( x: Int, y: Int) • Immutable • Instances are values • Simple!
- 38. case class Point( var x: Int, var y: Int) • Mutable • Instances are identities that track a series of Point states over time, where you only get to see the latest one • Not simple
- 39. "Classes should be immutable unless there's a very good reason to make them mutable....If a class cannot be made immutable, limit its mutability as much as possible." - Joshua Bloch Effective Java
- 40. “But the real world is mutable!” - Programmers
- 41. Entities vs value types “Some objects are not defined primarily by their attributes. They represent a thread of identity that runs through time and often across distinct representations.” - Eric Evans Domain Driven Design
- 42. Entities vs value types samuel_l_jackson.hairstyle samuel_l_jackson.occupation
- 43. Entities vs value types samuel_l_jackson.hairstyle samuel_l_jackson.occupation => “bald” => “actor”
- 44. Entities vs value types samuel_l_jackson.hairstyle samuel_l_jackson.occupation => “bald” => “actor” in 2017…
- 45. Entities vs value types samuel_l_jackson.hairstyle samuel_l_jackson.occupation => “afro” => “student” 1960s:
- 46. “Samuel L Jackson” H: “bald” O: “actor” H: “afro” O: “student” Identity 1960 1970 1980 1990 2000 2010
- 48. Immutability is just good programming!
- 49. Purity & types
- 50. Null is a 3rd world problem Cake cake = null; cake.cut();
- 51. Cake cake = null; cake.cut();
- 53. Sum types (Haskell) data Maybe a = Just a | Nothing
- 54. Sum types (Scala) sealed trait Maybe[+A] case class Just[A](a: A) extends Maybe[A] case object None extends Maybe[Nothing]
- 55. Sum types (Java) public class Maybe<A> { private Maybe() {} public abstract <B> B fold(Function<A,B> ifPresent, Supplier<B> ifAbsent); public static class Just<A> extends Maybe<A> { private final A a; public Just(A a) { this.a = a; } public <B> B fold(Function<A,B> ifPresent, Supplier<B> ifAbsent) { ifPresent.apply(a); } } public static class Nothing<A> extends Maybe<A> { public <B> B fold(Function<A,B> ifPresent, Supplier<B> ifAbsent) { ifAbsent.get(); } } }
- 57. public Session login( String userName, String pwd) throws AuthenticationException
- 58. public Session login( String userName, String pwd) throws AuthenticationException
- 59. public Either<AuthError, Session> login( String userName, String pwd)
- 61. Eradicable 3rd world diseases Nulls Exceptions Fall-through cases
- 62. What can we know from a type? process :: a -> (a -> Bool) -> WriterT Log (Reader Config) a • Takes a predicate that operates only on the given item • We do nothing with the “a” value other than filter it and return it • No side effects are performed • We might write an accumulating “Log” value • The code declares what will take place, but does not execute it when called • A “Config” will be required before the code can be eventually executed
- 63. Laziness
- 65. (define (sum-primes a b) (define (iter count accum) (cond ((> count b) accum) ((prime? count) (iter (+ count 1) (+ count accum))) (else (iter (+ count 1) accum)))) (iter a 0)) Sum primes, iteratively from SICP
- 66. (define (sum-primes a b) (define (iter count accum) (cond ((> count b) accum) ((prime? count) (iter (+ count 1) (+ count accum))) (else (iter (+ count 1) accum)))) (iter a 0)) Sum primes, iteratively Test each number to see if it is a prime
- 67. (define (sum-primes a b) (define (iter count accum) (cond ((> count b) accum) ((prime? count) (iter (+ count 1) (+ count accum))) (else (iter (+ count 1) accum)))) (iter a 0)) Sum primes, iteratively Accumulate total sum
- 68. (define (sum-primes a b) (define (iter count accum) (cond ((> count b) accum) ((prime? count) (iter (+ count 1) (+ count accum))) (else (iter (+ count 1) accum)))) (iter a 0)) Sum primes, iteratively Efficiency 👍 • The number range is only traversed once Modularity 👎 • Prime test and summation all smooshed together in loop
- 69. (define (sum-primes a b) (accumulate + 0 (filter prime? (enumerate-interval a b)))) Sum primes, with list ops Efficiency 👎 • The number list is traversed several times Modularity 👍 • Prime test impl & summation impl just plug in together
- 70. (define (sum-primes a b) (accumulate + 0 (stream-filter prime? (enumerate-stream-interval a b)))) Sum primes, with lazy stream ops Modularity 👍 • Prime test impl & summation impl just plug in together Efficiency 👍 • The number range is only traversed once
- 71. “Lazy evaluation is perhaps the most powerful tool for modularization in the functional programmer’s repertoire..” - John Hughes Why Functional Programming Matters
- 72. Living the dream Imagine a world where… You don’t have to chase down runtime stack traces in the logs, because they can’t happen
- 73. Living the dream Imagine a world where… Testing software is really easy
- 74. Living the dream Imagine a world where… You can write software by clicking typed functions together like Lego
- 75. Living the dream Imagine a world where… You can reason about one thing at a time without cramming the whole world in
- 76. User groups Melbourne Functional User Group https://www.meetup.com/en-AU/Melbourne-Functional-User- Group-MFUG/ Melbourne Scala User Group https://www.meetup.com/en-AU/Melbourne-Scala-User-Group/ Melbourne Haskell Users Group https://www.meetup.com/en-AU/Melbourne-Haskell-Users-Group/ clj-melb https://www.meetup.com/en-AU/clj-melb/
Editor's Notes
- #3: Powerful source of good software ideas Lambda functions Immutability Tuples Data classes Pattern matching Garbage Collection