Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hashing is a complicated affair ?

Author: Bob Durrett

Date: 09:28:57 04/06/04

Go up one level in this thread


On April 05, 2004 at 12:54:30, rasjid chan wrote:

>
<snip>
>
>Rasjid

After reading all of the bulletins following this one, it has become clear that
hashing is, indeed, "a complicated affair" and to some extent "everybody does it
differently."

All of the discussion has been about how to implement the hashing idea and the
overriding purpose seems to be to achieve greater overall computational
efficiency by avoiding unnecessary duplication of calculations.

As a user, and not a chess programmer, I see hash tables very differently.  To
the user, they are something the user may or may not be able to USE.

Post-mortem analyses are of greatest interest to me currently.  The user's
question for post-mortem analysis is: "how can I do a better [smarter] job of
post-mortem analysis by taking into account the way hashing works?  Clearly, if
every program does hashing differently, then the answer may be program
dependent.

Post-mortem analysis can be done in a variety of ways and the question therefore
becomes: "Which way is best?"  The user must think about how his/her actions
might be affecting the contents of hash tables and then choose the best
post-mortem analysis method.

The biggest [but not only] problem for the post-mortem analyst seems to be the
horizon effect.  The best way to deal with this problem may depend on how the
hashing is implemented.  From a user's viewpoint, the hash table contents is,
essentially, memory of analysis lines.  The user may be able to impact what is
put into that memory and in what order.

Users are in a "different world" from that of chess programmers!  : )

Bob D.



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.