Draft:Coalton (programming language): Difference between revisions

Revision as of 22:46, 16 July 2025

Revision as of 23:00, 16 July 2025

Line 33:

'''Coalton''' is a [[functional programming|functional]], [[static typing|statically typed]], [[General-purpose programming language|general-purpose]] [[programming language]] embedded in [[Common Lisp (programming language)|Common Lisp]]. Its type system is similar to Haskell's, but evaluation is strict like Standard ML or OCaml. [[Metaprogramming]] is supported through traditional Lisp [[Macro (computer science)|macros]]. Coalton focuses on performance through efficient data representation and compiler optimization<ref>https://www.youtube.com/watch?v=of92m4XNgrM</ref>.

''Coalton'' refers to both the language and its implementation, and remains under active, open-source development as of July 2025<ref>https://github.com/coalton-lang/coalton</ref>.

== History ==

Robert Smith createdstarted designing Coalton in 2018<ref>https://github.com/stylewarning/deprecated-coalton-prototype/blob/master/thoughts.md</ref> to conferimport the benefits of [[ML (programming language)|ML]]-like static typing withininto the interactive, incremental programming environment of Common Lisp. Coalton haswas sincefurther beendeveloped usedby fora quantumteam computingled researchby Smith and was officially announced in 2021<ref>https://coalton-lang.github.io/20211010-introducing-coalton/</ref>. Coalton has since been used for quantum computing research, the implementation of the [[Quil (instruction set architecture)|Quil]] compiler<ref>https://coalton-lang.github.io/20220906-quantum-compiler/</ref>, and defense applications.

Coalton is used<ref>https://meetings.aps.org/Meeting/MAR23/Session/F70.8</ref> at [[HRL Laboratories]] for building software for qubits based on exchange-only [[Silicon quantum dot|silicon dots]].

Line 44:

* Static typing with global type inference.

* AlgebraicParametric algebraic data types, pattern matching, and compile-time exhaustiveness checking.

* Multi-parameter type classes with functional dependencies.

* Machine code compilation (on Lisp hosts that support it).

* Compiler optimizations including heuristic inlining and user-controlled inlining with the <code>inline</code> directive.

* User-controllable function monomorphization with the <code>monomorphize</code> directive.

* Separate development and release modes.

* User-controller data type representation control with the <code>repr</code> directive.

* Separate development and release modes. Release mode enables greater compiler optimization opportunities at the expense of development interactivity.

* Non-allocating <code>Optional</code> data type (cf. Haskell's <code>Maybe</code> or OCaml's <code>Option</code>).

* Inline Common Lisp code with the <code>lisp</code> operator.

* Metaprogramming via ordinary <code>defmacro</code>.

* Advanced numerical data types including [[dual number|dual numbers]], hyper dual numbers, computable real numbers, and arbitrary precision floating point numbers.