On Apr 10, 2006, at 6:40 AM, David Zweigenhaft wrote:

Does anyone know how to achieve the following:
 

There's potentially an exponential number of these, assuming you ignore cycles. A naive algorithm would do a depth-first traversal from one of the nodes, recording paths along which the other node is found.

        Or, find the first N shortest paths between 2 nodes ?

One could use this to 
Excluding nodes is easy... you just don't allow the traversal to visit those nodes. Include constraints could get rather complicated quite quickly, although I imagine you might be able to start the traversal at one of the included nodes...
 
BWY - am I using the correct mailing list ?

Yes.

Doug