What I have in mind (next step)

[George Rudolph]

George 
> > However, consider this: It should now be relatively easy to constrain
> > the search space further by asking for solutions that contain specific
> > blocks, or (sub)sets of blocks, or to further constrain values of X
> > and/or Y.  Correct?
> >
> > That would yield one or more valid solutions in a smaller search space,
> > not enumerating every possible solution, but finding some, if they
> exist.
> >
> > And therefore getting a further glimpse into the structure of the
> > solution space.
> >
> > This is, I think, one of the strengths of a CLP approach as opposed to
> > other methods such as Combinatorial Optimization algorithms:  The
> > ability to add (or remove) recognizable constraints to tune the solver.
> 
Raph
> I don't know exactly what you have in mind, but that's certainly worth
> having a look at it.  An issue with the problem is the number of
> solutions.  Every trick that reduces the set of solutions to a smaller
> set that is representative of all solutions, is worth the pain.
>

What I intend here are things such as the following:

1. Add constraints to the existing code to say that 
one of the starter blocks must be {0, 12, 24} or
{0, 35,87}, or some set of two or more blocks.

2. Add constraints that specify that every block must follow some 
   additional desired pattern, such as {0, X, X+3}.