A concatenative programming language is one in which all terms denote functions and the juxtaposition of terms denotes function composition.[dead link][1] The combination of a compositional semantics with a syntax that mirrors such a semantics makes concatenative languages highly amenable to algebraic manipulation and formal analysis.[2]
Much of the original work on concatenative language theory was carried out by Manfred von Thun.
Properties
The properties of concatenative languages are the result of their compositional syntax and semantics:
- Concatenative languages are necessarily point-free as allowing terms to denote variables would violate the rule that all terms denote functions.
- The reduction of any expression is the simplification of one function to another function; it is never necessary to deal with the application of functions to objects.[3]
- Any subexpression can be replaced with a name that represents the same subexpression. This is referred to in the concatenative community as factoring and is used extensively to simplify programs into smaller parts.
- The syntax and semantics of concatenative languages form the algebraic structure of a monoid.[4]
- Concatenative languages can be made well-suited to an implementation inspired by linear logic where no garbage is ever generated.[5]
Implementations
The first concatenative programming language was Forth, although Joy was the first language to call itself concatenative. Other concatenative languages are Cat, Enchilada, Factor, Onyx, PostScript, RPL, Stappl, Trith and XY.
Most existing concatenative languages are stack-based; this is not a requirement and other models have been proposed. Concatenative languages are currently used for embedded, desktop, and web programming, as target languages, and for research purposes.
Most concatenative languages are dynamically typed. One exception is the statically typed Cat language which uses row polymorphism to assign types to functions that operate on stacks.[6]