Monadic Computation in Swift. Why?

Monadic Computation comes from Functional Programming which helps write composable functions while preserving failure context/side effects and provide better abstraction. Swift although not a functional programming language, can express monadic computations elegantly and can solve all the above-mentioned problems.

In the talk, I will show why do we need Monadic Computations and how we can implement in Swift. I will share real use case of monadic computations and how our team at Zalando is using this concept. There will be a short live coding to demonstrate that Monadic Computation can be expressed elegantly in Swift.

Transcript

1. TITLE IN CAPITAL LETTERS VIJAYA P. KANDEL @kandelvijaya iOS Engineer Why Monadic computation?
3. 3 When does it end? 1 more feature
4. 4 Spot a problem!
5. 5 Problems 1. Composability 2. Error propagation 3. True Type signature
6. 6 World full of failing cases (PROBLEM) Verify Monad propagates error and composes 2 Programming Paradigms Deducing answer from Monad Coding the Monad TABLE OF CONTENTS
7. 7 Are you connected? Is Wifi working? A world of 2 possibilities.
8. 8 How do we handle failable computation?
9. 9 The two halves Imperative Programming Functional Programming
10. 10 A simple game: MAD.A.DAM 1. M: Multiply input to a random Int 2. A: Add input to a random Int 3. D: Divide input by a random Int 4. Assume Swift doesn’t have Optional Perform: MAD.A.DAM on a dice roll input.
11. 11 A simple game: MAD.A.DAM
12. 12 A simple game: MAD.A.DAM
13. 13 A simple game: MAD.A.DAM
14. 14 A simple game: MAD.A.DAM
15. 15 A simple game: MAD.A.DAM
16. 16 A simple game: MAD.A.DAM
17. 17 The true division
18. 18 Contextual Box T Error Result<T> Contextual Box
19. 19 Contextual Box Result<Int> Div(a:) Int Mul(a:) Int Int
20. 20 MAD.A.DAM
21. 21 MAD.A.DAM
22. 22 MAD.A.DAM
23. 23 MAD.A.DAM
24. 24 MAD.A.DAM
25. 25 T Error T flatten()
26. 26 FLATTEN
27. 27 MAD.A.DAM
28. 28 RESULT MONAD
29. 29 RESULT MONAD T map flatMap T -> U U T -> U U Monad is a Contextual Type that provides linear binding and composition along with error propagation.
30. Monad is a CONTEXTUAL TYPE that provides easy BINDING/COMPOSITION without sacrificing Error Handling.
31. 31 RESULT MONAD T map flatMap T -> U U T -> U U Optional<T> is a monad. That is why we can chain stuff together. - map() - flatMap()
32. 32 RESULT MONAD T map flatMap T -> U U T -> U U Collection<T> is a monad. - map() - flatMap()
33. 33 RESULT MONAD T map flatMap T -> U U T -> U U Swift Try/Catch in principle can be implemented in terms of a Result Monad.
34. 34 MAD.A.DAM
35. 35 VERIFICATION
36. 36 Verification: True Types
37. 37 Verification: Error Propagation
38. 38 Verification: Error Propagation
39. 39 Verification: Composability Who cares error handling Solution! I Do! Monadic Solution!
40. 40 WAIT, What about the async task that might fail?
41. 41 Async Monad
42. 42 Async Monad
43. 43 Promise / Future / Deferred / IO :: Monad aTask: ( ( (T) -> Void)? -> Void )? T aTask T
44. 44 Promise / Future / Deferred / IO :: Monad aTask: ( ( (T) -> Void)? -> Void )? aTask T then/map T -> U aTask aTask T U T -> U
45. 45 Promise / Future / Deferred / IO :: Monad aTask: ( ( (T) -> Void)? -> Void )? aTask T bind/ flatMap T -> Promise<U> aTask aTask T U T -> U
46. 46 Promise / Future / Deferred / IO :: Monad aTask aTask T U Evaluate and finally call into outer (U -> Void) By intending to evaluate the inner block and calling into outer block with eventual value; we unpack a promise. None of the Pomise are evaluated unless someone calls `execute()` on the outermost layer. Thereby guarantee of order of execution is applied.
47. 47 Async Monad: USAGE
48. 48 Async Monad: USAGE
49. 49 Async Monad: USAGE
50. 50 Async Monad: USAGE
51. 51 Async Monad: USAGE
52. 52 Async Monad: USAGE
53. 53 Async Monad: USAGE
54. 54 A good programmer is someone who looks both ways before crossing a one-way street.” — Doug Linder
55. 55 References 1. Monad Research Paper http://homepages.inf.ed.ac.uk/wadler/papers/marktoberdorf/baast ad.pdf 2. Railway Oriented Programming https://fsharpforfunandprofit.com/rop/ 3. Dr. Bartosz Category theory lecture https://www.youtube.com/user/DrBartosz/videos 4. My blog for more detailed explanation https://kandelvijaya.com/2017/08/13/promisein30lines/ 5. Promise<T> from the talk https://github.com/kandelvijaya/MPromise 6. Result<T> from the talk https://github.com/kandelvijaya/ResultType
56. This presentation and its contents are strictly confidential. It may not, in whole or in part, be reproduced, redistributed, published or passed on to any other person by the recipient. The information in this presentation has not been independently verified. No representation or warranty, express or implied, is made as to the accuracy or completeness of the presentation and the information contained herein and no reliance should be placed on such information. No responsibility is accepted for any liability for any loss howsoever arising, directly or indirectly, from this presentation or its contents. DISCLAIMER 56
57. Vijaya P. Kandel vijaya.prakash.kandel@zalando.de @kandelvijaya 20-11-2017 Fashion Store iOS Engineer