The Multiparadigm Programming Language CCFL

Constraints support an efficient modeling and solution of problems in two ways: The most common application is their use for the description of problems with incomplete knowledge. On the other hand constraints may also serve as particular language constructs for the control of the program evaluation process. This paper presents the Concurrent Constraint Functional Language CCFL which comprises both aspects. Real-world problems comprise aspects from several realms. They are, thus, often best implemented by a combination of concepts from different paradigms. This combination is comfortably realised by multiparadigm programming languages, an area of research and application which has attracted increased interest in the recent years. The Concurrent Constraint Functional Language CCFL is a multiparadigm programming language which combines concepts and constructs from the functional and the constraintbased paradigms. The language enables the description of deterministic computations using a functional programming style and of non-deterministic behaviour based on constraints. Moreover, constraints are used to describe systems of concurrent cooperating processes and even typical parallelization patterns. We discuss language concepts and applications by means of examples. 1 Functional Programming In CCFL, functions are used to express deterministic computations. CCFL’s functional sublanguage inherits notions and concepts from the functional languages HASKELL and OPAL [Opa04, PH06]. It is a lazy language with polymorphic type system. A function consists of a type declaration and a definition allowing the typical constructs such as case-expressions, let-expressions, function application, and some predefined infix operator applications, constants, variables, and constructor terms. Free Variables In CCFL, expressions are allowed to contain free variables; this also applies to function applications. Function applications with free variables are evaluated using the residuation principle [Smo93], that is, function calls are suspended until the variables are bound to expressions such that a deterministic reduction is possible. For example, a function call (4 + x) with free variable x will suspend. In contrast, a com-