Anyone who is interested in Python and wants to explore new techniques. In this particular case using the constraint library, which uses a declarative approach: The problem is specified in a formalized way and no algorithmic code for Search and Constraint Propagation has to be written as it is contained in the CSP-solver.
Primarily it demonstrates how the Python constraint library can be used in an practical example such as solving a Sudoku puzzle. As there is already a lot of code for solving Sudoku puzzles, I based this example quite heavily on the great work of Peter Norvik and plugged-in the library parts at the appropriate places. The code is intentionally left close to Peter's source, so a diff shows the changed parts easily.
It is not optimized, so you will find that Peter's code is (sometimes quite a bit) faster, as obviously it has a Constraint Propagation optimized for Sudoku.