fun is the name of a simple functional programming language with fully lazy semantics and also the name of the program that reads a piece of source text in the language and evaluates it. The language is based on lambda-expressions, although thanks to the available extensive syntax most of the peculiarities of lambda-expressions are hidden from the user.
fun mimics the functionality found in mature functional programming languages like Miranda (TM), Haskell, and Gofer. Don't get confused: all built-in functions in fun are defined as prefix operators. fun does however support a crude facility for infix expressions: any prefix operator or user defined function may be enclosed in back-quotes (`) and as such used inbetween operands. Instead of back-quotes, a single dollar-sign in front of the operator name may be used as an alternative. All such infix operators, unlike the built-in ones, will assume the same (lowest) precedence and group from left-to-right (= left-associative), unless, of course, this default behaviour is explicitly overridden with the use of parentheses.
Missing in fun are function definitions by pattern-matching and any notion of modularisation, like modules or abstract data types. Also, fun is weakly typed and type-checking is done at run-time, e.g. lists may contain elements of any type. In contrast, Miranda requires lists to have elements of a single type, and supplies the fixed-sized tuples data type to handle record/struct-like data. For now, tuples may be emulated by lists and uninterpreted names can be used as tags.