Downloaded 25 times
Uploaded byPrateek Jain
Functional programming
This document discusses functional programming (FP) as an alternative to object-oriented programming (OOP). It notes issues with OOP like null pointer problems and inheritance issues. It then gives reasons for using FP like the need to write concurrent programs and handle data management problems. FP is described as using recursion instead of loops, higher-order functions, and treating programs as mathematical functions. FP avoids variables and assignments and favors immutability. Examples of FP languages provided are Erlang, Haskell and Clojure. Caution is given around maintainability, advocating for readability over one-liners. Real-world adoptions of FP at companies like Facebook and LinkedIn are also mentioned.
Functional programming
- 1.
- 2. Is OOPs, oops? •NullPointer problems • Inheritance issues • “new” operator issues
- 3. Is OOPs, oops? •NullPointer problems • Inheritance issues • “new” operator issues
- 4. Is OOPs, oops? •NullPointer problems • Inheritance issues • “new” operator issues
- 5. Why FP? • IHave to Be Good at Writing Concurrent Programs • Most Programs Are Just Data Management Problems (why ORM?) • More Modular • I Have to Work Faster and Faster • Code is written once, but read many times
- 6. Why FP? • IHave to Be Good at Writing Concurrent Programs • Most Programs Are Just Data Management Problems (why ORM?) • More Modular • I Have to Work Faster and Faster • Code is written once, but read many times
- 7. Why FP? • IHave to Be Good at Writing Concurrent Programs • Most Programs Are Just Data Management Problems (why ORM?) • More Modular • I Have to Work Faster and Faster • Code is written once, but read many times
- 8. Why FP? • IHave to Be Good at Writing Concurrent Programs • Most Programs Are Just Data Management Problems (why ORM?) • More Modular • I Have to Work Faster and Faster • Code is written once, but read many times
- 9. Why FP? • IHave to Be Good at Writing Concurrent Programs • Most Programs Are Just Data Management Problems (why ORM?) • More Modular • I Have to Work Faster and Faster • Code is written once, but read many times
- 10. What is FP? •RECURSION • ABSTRACTION • HIGHER ORDER FUNCTIONS • IMPACT MOST PROGRAMMING LANGUAGES
- 11. PROGRAMS AS FUNCTIONS •PROGRAM = DESCRIPTION OF A SPECIFIC COMPUTATION • Y = F(X) • F: X ->Y • MATHEMATICS • VARIABLES = ACTUAL VALUES • NO MEMEORY ALLOCATION CONCEPT • IMPERATIVE LANGUAGE • VARIABLES = MEMORY LOCATIONS + VALUES
- 12. PROGRAMS AS FUNCTIONS •NO LOOPS BUT RECURSION • NO VARIABLE EXCEPT AS A NAME FOR A VALUE • NO ASSIGNMENT OPERATION (x = x+1 , MEANINGLESS) • ONLY CONSTANTS, PARAMETERS AND VALUES
- 13. imperative functional EXAMPLE Void GCD (intu, int v, int* x) { Int y, t, z; z = u; y = v; While (y!=0) { … … } … } Void GCD(int u, int v) { if(v==0) return u; else return GCD(v, u%v); }
- 14. NO VARIABLES, NOASSIGNMENT • No notion of the internal state of a function. • Referential Transparency. • Value Semantics.
- 15. FP vs Others •Recursions instead of loops • Pattern matching instead of “if” • Pattern matching instead of state machines • Information transformation instead of sequence of tasks
- 16. FP vs Others •Persistent data structures • Powerful concurrency constructs : Actors • Software transactional memory • Avoid “Null”
- 17. What it reallymeans? • Immutability is good
- 18. What it reallymeans? • Immutability is good • No bugs (due to nasty side effects)
- 19. What it reallymeans? • Immutability is good • No bugs (due to nasty side effects) • Benefits from concurrency
- 20. What it reallymeans? • Immutability is good • No bugs (due to nasty side effects) • Benefits from concurrency • No more messy loops
- 21. What it reallymeans? • Immutability is good • No bugs (due to nasty side effects) • Benefits from concurrency • No more messy loops • Lazy evaluation
- 22. FP Examples • Erlang,Haskell, Clojure • F# • JAVA 8 (prject lambda), Scala, Groovy • R, Mathematica etc. (specialized languages)
- 23.
- 24. Maintaining, Maintainability • Usefunctional style wheretill it makes the intent more readable. Sort(extract(filter(having(on(Pet.class).getSpecies(), is(Dog)), pets, on(Pet.class).getName()), on(String.class)); List<Pet> dogs = filter(having(on(Pet.class).getSpecies(), is(Dog)), pets); List<String> dogNames = extract(dogs, on(Pet.class).getName()); List<String> sortedDogNames = sort(dogNames, on(String.class));
- 25. Maintaining, Maintainability • Oneliners are always not better. Convert(pets, new Convert<Pet, VetStay>() { @override public VetStay converter(Pet pet) { return new VetStay(pet, new Date(), “….”); }}); Private Converter<Pet, VetStay> toVetStay () { @override public VetStay converter(Pet pet) { return new VetStay(pet, new Date(), “….”); }}); Convert(pets, toVetStay());
- 26. FP - Adoption •Facebook (tchat), Linkedin uses Erlang • Twitter, UBS, Credit Suisse uses Scala
- 27.
- 28.