The promises of functional programming



This article has been written by Konrad Hinsen. It talks about the history of programming languages and explains why functional programming is different from traditional programming.
Programming has changed a lot and will continue to evolve: with all the new technologies that appeared over decades, new languages have been created (through assembly language to problem-oriented programming languages). Needs have evolved, programming languages have multiplied and programming approaches also: Functional programming has been growing in interest because it is considered more robust, compact and easily parallelizable (this means that computational tasks are divided into multiple communicating processes running in parallel) and testable. It is the realization of computations by composing functions (with input values and output values), but functional programming does not have variables because functions are not supposed to change any variable; and uses recursion instead of loops.

Functional programming uses algorithmic abstractions in the form of functions, which it considers as data. Functional programming is considered to be a promising language to deal with the issues of concurrency and parallelism although it must be said we are not yet at the stage of automatic parallelization due to compiler technology lagging in capabilities. 

There are different types of functional programming languages in terms of unavoidable side effects: the pure languages (no side effects: no mutation of local static variables or input-output flows for example) and the impure languages (use of side effects fully). Some example of functional languages: Lisp (impure programming language often used in artificial intelligence), the ML family (statically typed and impure), Haskell (a pure language). 





Commentaires

Enregistrer un commentaire

Posts les plus consultés de ce blog

Rich Hickey on Clojure

Rich Hickey is the creator of the Lisp programming language: Clojure (first version in 2007). The 22th of March 2010, he was interviewed on the Software engineering radio (radio dedicated to professional software developers). In this episode, Rich Hickey tells us that he did not start programming with Lisp, but with famous programming languages such as C++, C# and Java. Lisp is not new, it has been used in some domains successfully (in artificial intelligence for example) but don’t have a main stream. In respond of this, Rich Hickey says that Lisp was not designed as a main stream language and that it was designed for super users, researchers and very smart people who wanted to solve very hard and different kind of problems. Furthermore, programming has become like a social activity: More and more people are using Lisp, but nowadays most of the users are still scientists and researchers. However, for Rich Hickey: « every programmer with a good open mind will find working with t
Lire la suite

The roots of Lisp by Paul Graham

In this article, the author Paul Graham tells the history (“the roots”) of the programming language Lisp. John McCarthy has built Lisp in 1958 with only few simple operators and a notation for functions. The name Lisp is a mix of the words “List” and “Processing”. Full of mockery was made because of the use of many parentheses in the Lisp syntax, giving a funny acronym for LISP: "Lots of Irritating and Silly Parentheses". Because of its expressiveness and flexibility, Lisp had a lot of success in the artificial intelligence community. For the author, there are 2 unavoidable models nowadays: The C and Lisp programming languages, which influence the most recent languages (Most of the time, they are composed of a C base with adding Lisps parts, like runtime typing and garbage collection). Paul Graham also explains how John McCarthy obtained his new model of computation by creating functions ( eval) . Lisp is a language that does not distinguish between "expression
Lire la suite