Hi Danny,
happy new (civil) year. I'll respond to your message inline.
On Mon, 31 Dec 2012 21:04:12 -0000
"dannyjones183" <dannyjones183_at_yahoo.com> wrote:
> There are several sub-topics under discussion. I'm going to address them
> in multiple messages. First is Worrying Back ... and upgrades planned
> for my ephemeral solver.
>
> *) The definition of Worrying Back is more general than the constrained
> implementation described by Tom Holroyd in message #242 -- where a card
> from the foundation can be moved to an occupied column. This represents
> a restricted move environment that introduces a modest increase in the
> number of layouts that might need to be examined. An unrestrained
> implementation of Worrying Back allows a card to be moved from the
> foundation to an empty column as well.
Or an empty Freecell…
> This represents a multi-fold
> increase in the number of layouts that might need to be examined. It
> also represents a more complicated pruning algorithm and increased
> tracking of card movements.
Yes, I thought that will be the case.
>
> *) Michael Keller was interested in having a Worrying Back solution to
> (at least) one of the puzzles mentioned by Tom Holroyd. I wrote a
> quick-n-dirty patch to my standard solver that was designed explicitly
> to emulate Tom Holroyd's constraint on (only) the unsolved 8x4
> deals/games in the first 100,000,000 deals/games. The patch was never
> meant to be a functioning part of my standard solver!!! In particular,
> it was never meant to be part of the ephemeral solver because moves
> there are stored differently than in my standard solver.
>
> Bottom Line: It is not my intention to add Worring Back to any of my
> solvers as a permanent feature.
OK.
>
> Speaking of features, I hope to update my ephemeral solver to include
> command-line switches for: -KingOnly, -NoHome, -FourCell, and -Relaxed.
> These switches can be appended in any combination to the standard
> ephemeral command-line. However, this upgrade will be delayed until
> after I complete modifications that are under way to my standard
> solvers. I'm replacing multiple pruning features with a single pruning
> feature. I'm also upgrading the prioritization logic to assign a higher
> priority to standard moves equivalent to WKR automoves.
Nice, by the way are you using a version control system to maintain your
solver's source code? See
http://better-scm.shlomifish.org/ and
http://perl-begin.org/tutorials/bad-elements/#version_control (both of those
are my resources).
Furthermore, you may wish to look at the command line invocation of Freecell
Solver for some inspiration about flags:
*
http://fc-solve.shlomifish.org/docs/distro/USAGE.html (search for "Game
Variants Options")
Also see:
https://bitbucket.org/shlomif/fc-solve/src/b47591cad849a6f3398a4b3f9d3fe93aa6130eb1/fc-solve/source/scripts/gen_presets.pl?at=master
I should also note that according to UNIX conventions flags of more than one
character should start with two dashes instead of 1:
"--sequences-are-built-by=foo" or "--sequences-are-built-by foo" rather than
"-sequences-are-built-by foo". If you just do "-cas" it is equivalent to "-c -a
-s". Also see:
*
https://metacpan.org/module/Getopt::Long
*
http://www.gnu.org/software/libc/manual/html_node/Getopt.html
Unfortunately, due to stupid historical reasons, Freecell Solver does not follow
this convention, but you should not repeat my mistakes. :-).
Regards,
Shlomi Fish
--
-----------------------------------------------------------------
Shlomi Fish http://www.shlomifish.org/
Optimising Code for Speed - http://shlom.in/optimise
Chuck Norris does not code; when he sits at a computer, it just does whatever
he wants. — Kattana on Freenode’s #perl6 .
Please reply to list if it's a mailing list post - http://shlom.in/reply .
Received on Mon Dec 31 2012 - 15:22:40 IST