In a message dated 12/15/01 6:09:20 AM Pacific Standard Time,
aettlinger_at_worldnet.att.net writes:
> Just a quick reply to your latest response to my previous message. I
> had almost finished a much longer, well-considerd and brilliantly worded
> response when this laptop, which I really should abandon soon as my main
> mail environment, did it's nasty trick of spontaneously rebooting and
losing
> the entire draft, which I'd neglect to save after every paragraph, as I'd
> been trying to remember to do. So I now only have the patience to rewrite
a
> small sampling of my previous brilliant composition.
The Pulitzer Committee was disappointed. Can you (we) set some variable that
will automatically save each potential great work every, say, ten seconds?
> My first question, based on the fact that I'm virtually
> totally ignorant of what's "out there" is, what is really out there? For
> any future additions to FcPro, I think I'd want to make the rules that I'd
> turn over the source to any author who'd like to be incorporated, and
> he/she'd do the work of the implementation.
Let me suggest this. Leave FcPro just as it is. It's a pioneering work,
admirible for what it was able to accomplish in its original form with its
apparent straightforward simplicity (I've glanced at the organization of the
original.)
One stub for the original.
Keep stubs for variations on FcPro - your double hash, for instance, and
others' experiments.
Stubs for FCS and PatSolve. More stubs for the many more that are coming.
But FcPro may not be a good framework for all experimentation. Someone's good
idea may not fit neatly into Woods' organization.
Here are examples of what I mean. Dr. Tom tried my automove plan in PatSolve
(do I have the words correct?) yet the result surprised me with its mediocre
improvement and occasional lapse. My tactic didn't fit into his framework.
On the other hand, Dr. Tom shocked me by instantly testing my question which
involved bringing cards from the homecells back into the game. That would
have been a very difficult change to make in my own solver.
> If you want to take it from there, fine.
I wouldn't want to do anything with FcPro other than quickly analyze the
double hash tactic. As good as FcPro is, I must concentrate on my own solver.
BR
Received on Sun Dec 16 2001 - 08:37:15 IST