On Mon, 5 Sep 2005, Gary Campbell wrote:
>> -----------
>
> I'm not going to argue about what was above this line. As for the following....
>
>> I've tried running fcell using dosbox (a cross-platform DOS emulator) on my
>> Linux system. (after first changing user to a non-important account). After I
>> run "fcell 1" it prints the board layout and then say "Game #1" and then
>> prints four dots. Then I have to type "Enter" for it to print the solution.
>> Is that done on purpose? Or is there some sort of bug in dosbox?
>>
>> I tried solving a range, but no solution was displayed. Trying to solve game
>> #1941, (considered the hardest game) caused it to get stuck and nothing I've
>> tried released it. Pressing Ctrl+Break caused the cursor to stop blinking,
>> and dosbox to become un-closable. (I had to xkill it).
>
> I don't know about dosbox, when I run FCELL 1941, it takes about 3-4 seconds
> to display a couple of dots (indicating it is looking for shorter solutions), and gives
> me a result with 45 steps.
I see. Well, 3-4 seconds on native virtualization may run quite slowly
inside dosbox. I might try waiting a bit more.
> When I run FCELL * and enter 1941, it takes a fraction
> of a second to display a 46-step solution. BTW - when you solve a game and see
> dots printing out, that means it is in the process of finding shorter solutions. That
> can take a while on some games (not very many). After the first dot, you can press
> any key to see the best solution so far. You should not have to press a key to make
> it display the final solution, but, again, I don't know about dosbox.
>
I see. Well, the general convention for UNIX command line programs is that
once they accept arguments on the command line and read everything they
need from standard input, then they run and then display the output
immediately. They don't wait for user interaction. This is how FCS works
too.
In FCS, I solved the problem of trying to search for better solutions by
having the user specify a -opt flag in the command line, in case he wants
to. This used to make a relatively large optimization in the solution
length, but is no longer that dramatic due to the fact that now states
point to the original states from which they were discovered.
>
>> BTW, have you given any consideration in regards to the license of the solver?
>> Or is it just Gary
>> Campbell's-extremely-restrictive-license-use-at-your-own-risk-this
>> -solver-will-eat-your-dog?
>>
>> From my download page on my website, click the "little man" and you'll go to
> my home page. That explains my copyright. I ask for no license.
>
You mean:
<<<<
Although protected by Copyright, I grant permission to copy my articles
and software, but only if not-for-profit, and only if they include my
Title and Copyright, and a reference to this website. Notification to me
and courtesy copies of the publication would be appreciated. For other
publication, please contact me directly.
>>>>
By license I meant something along the lines of The GNU General Public
License (or GPL), the BSD license, the MIT X11 license, LGPL, MPL, CPL,
Sleepy Cat, etc. etc. These are legal documents that govern what can be
done with the program and what is forbidden. Your informal licensing terms
would be considered quite problematic in the UNIX/Free and Open Source
Software world.
A final note is that I linked to your solver as well as to this discussion
from the Freecell Solver homepage:
http://vipe.technion.ac.il/~shlomif/freecell-solver/links.html#other_solvers
It's near the bottom of the section. Please read it and let me know if
it represents your solver well.
Regards,
Shlomi Fish
----------------------------------------------------------------------
Shlomi Fish shlomif_at_vipe.technion.ac.il
Home Page:
http://www.shlomifish.org/
- "Sometimes two bugs cancel each other."
- "It wasn't a bug - it was a sanity check!"
Received on Thu Sep 08 2005 - 14:13:05 IDT