Constraints in Procedural and Concurrent Languages

This chapter addresses the integration of constraints and s earch into programming languages from three different points of views. It first focuses on the use of constraints to model combinatorial optimization problem and to easily implement search procedures, then it considers the use of constraints for supporting conc urrent computations and finally turns to the use of constraints to enable open implementations of constraints solvers. The idea of approaching hard combinatorial optimization problems through a combination of search and constraint solving appeared first in log ic programming. The genesis and growth of constraint programming within logic programming is not surprising as it catered to two fundamental needs: a declarative style and non-determinism. Despite the continued support of logic programming for constraint programmers, research efforts were initiated to import constraint technol ogies into other paradigms (in particular procedural and object-oriented paradigms) to cate r to a broader audience and leverage constraint-based techniques in novel areas. The first mo tivation behind a transition is a desire to ease the adoption of a successful technology. Moving constraints to a platform and paradigm widely accepted would facilitate their adoption within existing software systems by reducing resistance to the introduction of technologies and tools perceived as radically different. A second motivation is that constraints are vers atile abstractions equally well suited for modeling and supporting concurrency. In particular, concurrent computation can be seen as agents that communicate and coordinate through a shared constraint store. Third, constraint-based techniques can leverage the semantics of a target application domain to design specialized constraints or search procedures that are more effective than off-the-shelves constraints. The ability, for domain spec ialists, to easily create, customize and extend both constraints solvers and search is therefore a necessity for adaptability. The continued success and growth of constraints depends on the availability of flexible, extensible, versatile and easy to use solvers. It is conting ent on retaining performance that