Author: Rémi Coulom
Date: 12:49:51 09/15/98
Go up one level in this thread
Hi, Here are my thoughts about what you suggest : The first thing to do would be to design a low-level protocol for exchanging messages on the serial link. Chrilly Donninger already did this with auto232 and many DOS & Windows chess programs support it. So I think any new autoplayer interface should be upward-compatible with auto232 low-level communication protocol on the serial link so that unix engines could play against Genius, Fritz, Rebel, etc. Once the goals and specification of the low-level layer is defined, building a driver to connect a text-based chess engine via pipes should be relatively easy (a bit like the noname driver for DOS programs). I have very recently implemented Windows autoplayer support for TCB (in version 0038) and installed Linux on my PC. Maybe the first thing to do would be to make a linux auto232 compatible program that does not make any extensions to Donninger's interface. I think I could do this in TCB relatively easily when I have understood how to do i/o on the serial link in Unix (maybe multithread is needed also to do like Stefan Meyer-Kahlen does ?), and publish the source code so that people can test it in Unix. By the way, I corrected two bugs in the code distributed by Chrilly Donninger for Windows Autoplayer ( version of 08.07.1997) : extended mode does not work and there is an error in the game number transmitted to the slave program when saving a game. The corrected source code will be published on my web page for the next release of TCB (0039). Rémi
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.