: Constraint satisfaction programming (CSP) is a powerful paradigm for solving complex combinatorial problems, which has gained a lot of attention recently. Putting the power of constraint satisfaction into the hands of programmers and designers in a simple fashion is, however, still an open issue. We propose to classify the various approaches to this problem into three categories: the language approach, in which only basic mechanisms are proposed to the user, the library approach, in which the user may pick up pre-designed algorithms off-the-shelf, and the object-oriented framework approach, an intermediary position between languages and libraries. We outline the design of such a framework and stress on the various advantages of this approach compared to the other ones.