Download to read offline
More Related Content
![[4DEV] Bartosz Sokół - Functional developer in object oriented world - how F#...](https://cdn.slidesharecdn.com/ss_thumbnails/bartoszsokol-functionaldeveloperinobjectorientedworld-170405111747-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to Functional programming
Functional programming
- 1.
- 2. Objective The popularity offunctional programming has been increasing lately. Specially pure functional languages like F#, Haskell, Scala, Erlang and Elixir are becoming more popular. Even OOP languages like C# and Java have included some functional libraries Ex: LINQ and lambda functions This session will focus on what Functional Programming is, how it is implemented and why it has become popular.
- 3. Content Programming Paradigms Imperative vs.Declarative Programming Functional Programming Benefits of Functional Programming Main Concepts ◦ Pure functions ◦ Immutability ◦ Higher Order Functions ◦ Referential Transparency Implementing functional programming with C# Multi-paradigm approach Popularity of Functional Programming
- 4. Programming Paradigms A programming paradigmis a way of programming It is a style of approaching a programming problem. Types ◦ Imperative ◦ Declarative
- 5. Imperative Programming Focus on'How' something needs to be done Explicitly declares sequence of steps to be followed that changes the state Types ◦ Procedural ◦ Object Oriented Programming
- 6.
- 7. Declarative Programming Focuses on'What' the result should be Expresses the logic of the program without explicitly declaring the steps Types ◦ Functional Programming ◦ Logic Programming ◦ Database approach
- 8.
- 9. Issues in ImperativeProgramming Less support for Concurrency Vulnerable to 'Race Conditions'
- 10. Functional Programming is oneway to address these issues effectively
- 11. What is Functional Programming? "In computerscience, functional programming is a programming paradigm—a style of building the structure and elements of computer programs—that treats computation as the evaluation of mathematical functions and avoids changing-state and mutable data." -Wikipedia
- 12. / Functional Programming Treats functions asFirst Class values Avoid state mutations
- 13. First Class Values Treat functionsas values Pass functions as data • Be refered from constants and variables • Be passed as a parameter to another function • Be returned as a result from another function Functions as First-Class Entities can:
- 14. Avoiding State Mutation Avoid destructiveupdates The value stored prior to the update should not be destroyed
- 15. Why Functional Programming? LESSBUGS SUPPORT FOR PARALLEL PROGRAMMING (THREAD SAFE) EFFICIENCY
- 16. Other Benefits INCREASED READABILITY INCREASED UNDERSTANDABILITY EASIER FORTESTING AND DEBUGGING
- 17. Main Concepts PURE FUNCTIONS RECURSION REFERENTIAL TRANSPARENCY HIGHERORDER FUNCTIONS IMMUTABLE VARIABLES
- 18. Pure Functions A purefunction: ◦ Returns the same result if given the same arguments ◦ Causes no observable side effects
- 19. Returning same resultfor same arguments The value of the variable on the global context has changed
- 20. Ex: Functions thatread files or generate random numbers can never be pure
- 21. No Side Effects SideEffect : ◦ An interaction with an external mutable state apart from the function's input and output A side effect can be : ◦ Mutation of a global scope ◦ Mutation of the input arguments of a function ◦ Throwing exceptions ◦ Performing any I/O operation
- 24. Function Honesty A conceptsimilar to pure functions The function is an 'honest' representation of its method signature The behavior can be predicted using the signature.
- 25. Recursion Recursion is whena function calls itself repeatedly. Used to remove local side effects Specially in loops and similar iterative structures
- 26. Example: Sum ofintegers from 1 -10 Non-recursive method Recursive method Interacts with external variable result.
- 27. Referential Transparency Being able toreplace a function with its result. Supported by Pure Functions and Immutable Variables ◦ Pure functions give the same result for the same arguments
- 28. Higher Order Functions Functionsthat: ◦ Take one or more functions as arguments AND/OR ◦ Returns a function as a result
- 30. Immutable Variables Variables andobjects that once created do not change Always maintain the same state Thread safe Readability increases
- 31.
- 32.
- 33. Functional Programming withC# C# mainly uses an imperative programming approach (It's basically an OOP language) But it also has some declarative approaches Ex: LINQ library, lambda functions
- 34. FP Vs. OOP FunctionalProgramming OOP Use Immutable data Use Mutable data Declarative programming Imperative programming Support parallel programming Less suitable for parallel programming No side-effects Can produce side-effects Flow control using function calls and recursion Flow control using loops and conditional statements Iterate collection data using recursion Iterate collection data using loop Execution order of statements is less important Execution order of statements is important
- 35.
- 36. LINQ LINQ is afunctional library Follows declarative approach for common operations on sequences (list) Ex: Map Filter Sort
- 37. LINQ -Map Getting anew sequence from a given sequence where the elements of the new sequence have been obtained by subjecting to a function Example:
- 38. LINQ - Filter Gettinga new sequence from an old sequence where the elements of the new sequence are those that pass a certain condition (predicate) Example:
- 39. LINQ - Sort Gettinga new sequence by ordering the old sequence according to the key provided in a key selector function. Example:
- 40. Method chaining Method chainingis a way of calling a method directly onto the result of a previous method. We do this all the time with LINQ. It is a functional approach because it avoids temporary variables and state changes.
- 41. Fluent Interfaces Fluent interfacesare an implementation of method chaining. It is a more complex version of the simple method chaining. The main feature of fluent interfaces over method chaining is that it is self referencing. Not all implementations of method chaining will be a fluent interface The Where and OrderBy functions act on the same instance of the IEnumerable returned from the Select function.
- 42. Function Honesty An honestfunction is a function that does exactly what it says in the function signature. No unexpected outputs or inputs. The exception is not an output predicted by the function signature. Then this is a dishonest function
- 43. This method signaturesays that this function will definitely return an integer. But actually the truth is this function might return an integer. It can also throw an exception.
- 44. Functional Error Handling Imperativeapproach ◦ throw/try/catch -> Side effects Declarative approach ◦ Treat the error as part of the payload Approaches: ◦ Implementation of Option or Either types
- 45. Is C# apure functional language? The answer is both YES and NO. NO because C# is mainly a OOP language and follows imperative style YES because C# has some support for functional concepts like with LINQ C# is a Multi-Paradigm language
- 46. Multi-Paradigm Approach Actually themulti-paradigm approach is a lot more practical when considering large programs rather than the pure functional approach. It's very difficult to apply functional approaches to an entire program Ex: I/O operations are considered side effects that need to be avoided in FP. But in a real world application it is impossible to not have I/O operations at all. So, the solution is to isolate the I/O operations to be performed imperatively and have everything else follow a functional approach
- 47. INPUT OPERATION SOME OTHER OPERATIONSOUTPUT OPERATION Do this functionally. There can be no side effect causing functions here These should be isolated. You can use an imperative approach here Example:
- 48. Multi-paradigm approach can alsohelp avoid some of the other issues in pure functional programming
- 49. Problems with purefunctional programming LEARNING CURVE DIFFICULT TO WRITE COMPLETE PROGRAMS IMPLEMENTING ONLY FUNCTIONAL PROGRAMMING HIGH MEMORY CONSUMPTION NEED TO WRITE EXTRA CODE
- 50. Why is FP trendingright now? Pure functional languages like F#, Haskell, Erlang and Elixir are becoming more popular recently. This is mainly because of the greater support for concurrency and parallel processing. A lot of other traditionally non-functional languages and frameworks have also introduced different functional utility libraries to support functional programming. Ex: RxJs, lazy.js, immutable.js and rambda
- 51. Success Stories WhatsApp supports900 million users with only 50 Engineers using Erlang Discord handles over a million requests per minute using Elixir Facebook's anti-spam system uses Haskell For each system the most important aspects were concurrency and speed. They were able to achieve it using functional programming
- 52. Recap There are 2main programming paradigms, Imperative and Declarative Functional programming is a declarative approach that avoids state changes and mutable data The main concepts are Pure functions, recursion, higher order functions, immutability and referential transparency. C# is a multi-paradigm language that has support for functional programming approaches. Ex: LINQ library The main problem with Functional Programming is that it is difficult to implement to a whole program. But it is more popular now because of greater support and suitability for parallel processing
- 53. Resources Functional Programming inC# by Enrico Buonanno .NET documentation
- 54.
- 55.
Editor's Notes
- #27 Crop