The practical benefits of constraint propagation really began to emerge when it was embedded in a programming language
(Van Hentenryck, 1989). Again it was logic programming that was first used as a host language, producing impressive results first on some difficult puzzles and then on industrial problems such as scheduling, warehouse location cutting stock and so on
(Dincbas et.al., 1988). The embedding suggested new kinds of propagation, new ways of interleaving propagation and search and new ways of varying the propagation according to the particular features of the problem at hand.
These advantages were very clearly illustrated when, using lessons learned from the Operations Research community, constraint logic programming began to outperform specialised algorithms on a variety of benchmark problems. However the main advantage of constraint programming is not the good performance that can be obtained on benchmarks, but its flexibility in modelling and efficiently solving complex problems.
A constraint program for an application such as Vehicle Scheduling, or Bin Packing, not only admits the standard constraints typically associated with that class of problems, but it also admits other side-constraints which cause severe headaches for Operations Research approaches.
Currently several companies are offering constraint programming tools and constraint programming solutions for complex industrial applications. As host programming language, not only logic programming but also LISP and C++ are offered.