On Friday 14 January 2005 22:33, daj95376 wrote:
> Your recent posting broaches the question of *where next?* for
> FreeCell software development.
>
> In the past, I believe the primary emphasis was on testing a broad
> range of deals to see which were solvable or not. Fast solvers like
> FreeCell Solver and Patsolve were able to provide the answers.
>
> Now, I believe that FreeCell solvers need to change their emphasis and
> address the needs of those who are still playing the game.
>
> There are three primary types of players that come to mind: (1) those
> who want to see a meaningful solution to a deal that's giving them
> trouble, (2) people who are looking for comparably difficult deals to
> tackle next, and (3) those who want to know if a solution exists from
> their current position.
There are also those who wish to find as many interesting solutions as
possible to a certain deal.
>
> These people are often less interested in *how fast* a deal can be
> solved -- just so long as they don't get impatient while waiting for
> results -- than the quality of the results they receive.
>
> FcPro26 addresses the group (3) MS FreeCell players. NetCell addresses
> the group (2) non-MS FreeCell players. But, I don't know of any
> FreeCell solvers that address the group (1) players!
>
I believe any solver that can solve the board is good for them.
> If someone's looking for a solution to MS deal 11575, I think they'd
> prefer to have a reasonably intuitive 64-move solution to examine
> instead of a convoluted 169-move solution -- even if they can get the
> second solution faster.
Yes, but if you have to wait a few hours for the 64-move solution, while a
somewhat longer solution can be produced in a few seconds, then obviously the
latter is preferable. I took some measures to make the solutions produced by
Freecell Solver become shorter, and I don't think that they are too long as
they are.
Regards,
Shlomi Fish
>
> It's possible that no one other than myself wants any of what I've
> indicated. However, that's the direction I'm taking with my MS
> FreeCell solver. Currently, it generates an average of 50.25 moves (64
> moves max) for the 32K MS deals and I find playing a solution to often
> be informative; i.e, group (1). I'm striving to make its moves
> efficient yet *natural* where possible. I'm also working to improve
> closing moves without penalizing initial and intermediate moves. I
> haven't made much progress towards group (2) except to find deals with
> trivial solutions -- of which there are many. My solver isn't
> interactive, so it isn't of any help for group (3).
>
> Regards, Danny A Jones
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
--
---------------------------------------------------------------------
Shlomi Fish shlomif_at_iglu.org.il
Homepage: http://www.shlomifish.org/
Knuth is not God! It took him two days to build the Roman Empire.
Received on Sat Jan 15 2005 - 11:42:42 IST