On Wed, 19 Dec 2001 WKRfresno_at_aol.com wrote:
> > This would be more machine readable if it were one move per line.
>
> One move per line does not make a printed solution more machine
> readable..
Just simpler. Yes, almost anything is machine readable.
> In my ideal world there would be just two formats: concise machine readable,
> human readable. Every solver would be able to read the first one, produce
> both.
So what's really needed is a spec for the concise format. Naturally we
don't want to do any work on the output side of our programs, :-) so the
spec should be flexible enough to accomodate a variety of possibilities.
A standard reader with hooks for outputting different "moves" in internal
formats might be written.
Of course some human will eventually copy&paste the human readable
solution off a web page and pass it to the solver...
> If MK needs all solutions on his site to be concise for reasons of
> space, he could provide the human readable form of a single solution
> in response to a click.
That could work.
> > 6S to 7D
> > 2S to 3H
> > 5H to 6S
> > AD out
>
> Now I see how you avoid using column numbers. Using column numbers appears to
> be necessary if sequence moves are allowed.
Why? When I say 6S to 7D it _could_ mean that I'm moving a pile, if the
6S has cards on it. The only time this doesn't work is when you're using
multiple decks. A concise, universal format would be more useful if it
could handle any of the hundreds of variants of these games, including
the multiple deck ones. Shlomi? Leaving out the games with hidden cards,
which are naturally impossible to specify to a solver, what are the kinds
of moves that need to be handled?
Received on Wed Dec 19 2001 - 18:51:23 IST