From: Peter Burns <rictic@...>
Date: 2006-01-01T09:24:02+09:00
Subject: Re: [QUIZ] Numeric Maze (#60)

On 12/31/05, Stephen Waits <steve@waits.net> wrote:
> Since that's about all I'm going to be able to do on this quiz, here
> are some of my test cases.  I'm pretty confident in the shorter
> solutions, but have less confidence in the longer ones.  These were
> all discovered via stochastic search.
>
> I hope this won't qualify as "code", as mentioned above - since the
> idea is that everyone may test their solutions against this, the fact
> that it's already coded as test cases seems naturally acceptable to me.
>
> I'm very much looking forward to the solutions on this one!
>
> For now, I've posted them on http://swaits.com/
>
> --Steve

I think I've improved upon this test code a bit.  I found this one to
be a bit brittle, as it will fail solutions with unanticipated paths
of the same or smaller length.  I lost some of the metadata in the
switch, as Steve has the different solutions ordered by type (i.e.
primes, powers of two, etc).

http://rafb.net/paste/results/qwhnUg32.nln.html

Any feedback is, of course, appreciated.

I'm just pounding away at this test case code because my mathematician
buddy is busy at the moment.  This is quite the nontrivial problem...