Author: Rémi Coulom
Date: 09:26:30 09/24/01
Go up one level in this thread
On September 24, 2001 at 10:43:02, Georg v. Zimmermann wrote:
>Hi !
>
>I searched in the last few pages but can't find your original posting anymore,
>can u repost or send to d2d4@d2d4.de ?
>
>Because I don't understand how the pretty picture relates to a search tree.
>
I do not know how to find old messages either. Just take a look at:
http://www.cs.umd.edu/hcil/treemaps/
What I use is actually a variation of the "squarified treemap" algorithm
described in this paper:
http://www.win.tue.nl/~vanwijk/stm.pdf
(my algorithm does not sort nodes by size, so that move ordering information can
be visualized too: First node is at the top left, last node at the bottom right)
>How will I store the search tree in pgn ? Will I be able to add more info like
>which extensions I did, what my alpha and beta values where, why this tree was
>razored, whether the move came from the hash or was ordered the first ?
>
I will write a detailed documentation when the program is out. PGN allows to
store games with variations and this is how the tree is represented. Flags such
as "fail-high", "move from hash", ... can be set by giving this signification to
bits of the numeric annotation glyph (check the PGN standard for details). My
program can also read comments and show them. For instance, The Crazy Bishop
writes {alpha value beta} in comments.
>Visualizing the search tree helps a whole lot in finding bugs etc, thanks for
>releasing this !
>
>Georg
>
Check
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.