Functional programming for the Advanced Beginner
Download as PPTX, PDF3 likes806 views
Functional Programming for the advanced beginner covers functional programming concepts like composition, currying, functors, applicatives, and monads. It provides an overview of functional programming principles like avoiding side effects, favoring expressions over statements, and immutable data. Key topics include functional composition by combining functions, partial function application using currying, and using functors, applicatives and monads to manage effects in a declarative way. The document emphasizes that functional programming is about mapping types rather than mutating state, and controlling program flow through function composition rather than conditionals or loops.
Functional programming for the Advanced Beginner
- 1. Functional Programming for the advanced beginner By Luis Atencio (@luijar) https://joind.in/talk/d4540
- 2. About me (@luijar) • Author of: • Functional Programming in JS • RxJS in Action • UXDevSummit2018: • Workshop on FP • Videos: • https://www.udemy.com/functional-php-7/
- 3. Outline • Beginner • FP Primer • Composition • Curry • Advanced • Functors • Applicatives • Monads • Recursion & pattern matching • Dealing with effects
- 5. “Not Awesome: ES6 Classes” -Josh Burgess
- 6. OOP is dead ONLY if FP is dead
- 7. Paradigm subtraction (abstraction) Paradigm Mass adoption Subtraction Modular 1970s Limitless files Structured 1980s Removes “goto” Object Oriented 1990s Removes function pointers for polymorphism Functional ? Takes away assignment and mutable state
- 8. A Primer
- 9. Background • Lambda calculus • Category Theory • Categorical properties: identity, composition, and morphisms (functions) • Algebraic data types: functors, applicatives, monads
- 10. Basic Principles • A function is a mapping from one type (A) to another (B) within the same category of objects • f :: A -> B • Everything is a function • Functions are the main unit of work • Functions are first-class citizens • Functions are of higher-order • Favor expressions vs statements • Avoid externally visible side effects • Avoid mutable state
- 11. Avoid side effects • Keep changes from leaking • If you must, move them to the boundaries of your logic • Manage effects with declarative, lazy containers (e.g. IO monad)
- 13. Things you’ll hardly see in FP • Counted loops or while statements • If-else conditions (replaced by pattern-matching) • Ternary operator accepted <condition> ? <expression> : <expression> • Exceptions thrown • Global variables passed around and/or referenced within functions • Long functions (5+ lines) • Void functions • Variable re-assignment (e.g count += i )
- 14. Composition
- 15. “Simple ideas combine to form complex ideas” -John Locke
- 16. Composition: control flow in FP
- 17. Composition of 2 functions
- 18. Example: counting words in a file
- 19. Example: reading words from a file2 In: string Out: number
- 22. Curry
- 23. Currying is essential for composition • Partially evaluate functions • Never evaluate a function with missing data • Transform multi-arity in to single-arity functions • Functional languages use auto-currying + lazy evaluation
- 24. Simple example
- 25. Sequences of partial function calls
- 26. Build abstractions over APIs OR
- 27. Functors
- 28. What are Functors • Functions are mappings from one object to another • Functors are mappings from one category to another • Most common is the endofunctor (same categories) • Functors define a function named map (fmap)
- 29. What are Functors • Functions are mappings from one type to another • Functors are mappings from one context to another • Most common is the endofunctor (same context) • Functors define a function named map (fmap)
- 31. Maybe functor Rudimentary pattern-matching
- 32. Workings of Maybe Functor
- 33. Functor/Composition equivalence • Define map for function types • Equivalence: map can be derived from compose
- 34. Applicatives
- 35. What is an Applicative (functor) • Functor with extra laws and operations • Not just wrap values, wrap functions too! • Implements the Apply interface: ap
- 37. Workings of an applicative
- 39. Monads Applicative + Functor + Effects
- 40. What is a monad • Functors apply a function to a wrapped value • Applicatives apply a wrapped function to a wrapped value • Monads apply a function that returns a wrapped value to another wrapped value. Enter flatMap (bind, chain)
- 41. Maybe Example: object traversal
- 42. Maybe Example: object traversal2
- 43. Maybe Example3
- 44. Back to composition Point-free coding
- 46. A few more monads in the wild • Algebraic Data Types: • Either, Maybe, Try, Validation (sum type) • Tuples, Records, State (product type) • State management • Do, State, IO • Async • Promise • Observable • Dependency Injection • Reader • Many more...
- 48. Imperative code
- 49. List Sum
- 50. List insert
- 52. IO: Monads as computations with effects Keep effects at the boundaries Keep effects at the boundaries
- 53. Parting Thoughts • FP makes code easier to reason about • FP is closer to the maths and propositional logic • No need for patterns!
Editor's Notes
- #5: Use Ramda to bring FP primitives into the language JS has lambdas and higher order functions JS has support for objects and classes (limited)
- #6: https://medium.com/@dan_abramov/how-to-use-classes-and-sleep-at-night-9af8de78ccb4 https://github.com/joshburgess/not-awesome-es6-classes/
- #13: Endofunction -> a function with equal domain and codomain
- #22: Millennium Facon
- #30: Slide switches “category” for “context”
- #32: Used for error handling