Abstract:
When more that two decades ago David Waltz presented his now well-known filtering algorithm for labeling three-dimensional line-diagrams, one could hardly expect that the basic principle he introduced, namely the technique of constraint satisfaction, would become the basis of an emerging research area which has already outgrown the field of artificial intelligence and started to influence many other disciplines. In the last few years many AI problems have been formulated as constraint satisfaction systems. Usually such a system offers a proper syntax for specifying constraint satisfaction problems. However, the syntax sometimes may be complex or awkward, and may require some effort to get familiar with. In this paper, we argue that it is more appropriate to specify a constraint satisfaction problem by using a graphic editor and to translate the output of the editor into a notation that can be used as input for the constraint satisfaction system. Moreover, we are aiming at using the same output for documentation purposes. In other words, we are applying Knuth's idea of literate programming to constraint satisfaction.