Knight's tour Warndorff's algorithm problem

[Robin Rytich]

On Wed, 2010-03-10 at 02:37 -0300, Gabriel Genellina wrote:

> Warnsdorff's algorithm is heuristic; it works most of the time, but in  
> some cases leads to a dead end and you have to backtrack and try another  
> alternative.
> The starting square is important; if you start at 1,1 (instead of 0,0)  
> your program finds a solution for all those problematic board sizes.

Wow, didn't know about that. It seems to be a good idea for me to make a
little research around this question.

> Some comments on the code itself:

> This sets a class attribute (as opposed to normal, instance attributes)  
> which is shared by all ChessBoard instances (this really should be covered  
> in the FAQ!). 

Damn, I'm ashamed.

About other points, I actually totally agree with you (and corrected
almost everything before received your letter). Thank you for your
remarks, I'll review public code more careful next time.

Robin Rytich