What is Functional Programming?

Functional programming is presented as a style where program logic is expressed primarily through function application and composition. The article argues that this usually depends on higher-order functions, functions as first-class values, expressions instead of statements, immutable data, and a reduced use of effects.

Reading notes

#

• A strict definition of FP as programming with pure, side-effect-free functions is described as too narrow for common usage.
• In collection processing, FP is shown through declarative operations such as map, filter, and reduce.
• Functions can be treated as values, which avoids duplication and makes composition easier.
• Immutable data structures support updating by creating new values instead of mutating existing ones.
• Effects can be represented as values, separating the description of a computation from its evaluation.
• Errors can also be modeled as values, such as Result or Either, instead of being handled through exceptions.
• FP often separates data from behavior, while OO tends to combine them through encapsulation.
• Algebraic data types and pattern matching are presented as core building blocks of functional programs.
• The article distinguishes casual FP, where functions may have effects, from pure FP, where all functions are effect-free.
• Purity is described as valuable but not something most mainstream languages can fully verify.
• The practical value of FP is said to come mostly from moving away from imperative style toward casual FP.
• Languages differ in how well they support FP ergonomically, with Java and C# on one end, Haskell on the other, and Kotlin or Rust in between.
• The article concludes that FP is useful as a set of techniques, not as an end in itself.