Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: A chess program in SQL

Author: Stefan Meyer-Kahlen

Date: 01:37:53 05/19/00

Go up one level in this thread


On May 19, 2000 at 02:18:17, Dann Corbit wrote:

>On May 19, 2000 at 02:01:48, Stefan Meyer-Kahlen wrote:
>
>>On May 18, 2000 at 15:24:42, Olaf Jenkner wrote:
>>
>>>>
>>>>It would be interesting to run it in our SQL lab :-)
>>>>
>>>>Eugene
>>>
>>>It would be interesting to write a chess program in SQL even
>>>when it will be very slow.
>>>I wrote a SQL-procedure to solve the N-Queens-Problem. It was a
>>>funny work.
>>>
>>>OJe
>>
>>Good luck and have fun :-)
>
>I think you could actually make a program that plays very interesting chess.
>The reason I say that is if you build a crafty book with 2.3 million chess games
>and don't throw out a single move (yes, I am deranged) it will stay in book for
>an incredibly long time.  Basically, all it is doing is a database lookup.  You
>could easily do the same thing in SQL.
>
>Some databases like Rdb and DB2 allow foreign stored procedures in C, Fortran,
>or whatever.  With a getup like that you might actually create a chess program
>superior to anything else because you could easily and dynamically alter the
>behavior and performance of the book by altering the SQL queries to find moves.
>
>For instance, you might start off with games played by 2600+ super GM's.  And if
>you get in trouble, seek any line that still has wins in it.  With 120 million
>distinct board positions, you can stay in book sometimes clear to the end.

I don't think that using such a big book will help you much more compared to a
"normal" book with a couple of selected games (maybe 300.000) stored only to a
certain ply.

Stefan



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.