Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: new autoplayer interface - software engineering approach

Author: Rémi Coulom

Date: 11:26:29 09/17/98

Go up one level in this thread



On September 16, 1998 at 21:48:57, Robert Hyatt wrote:

>On September 16, 1998 at 15:11:18, Moritz Berger wrote:
>
>>Hi Bob,
>>
>>why don't have an interface connection layer (engine <-> ICL <-> device) that
>>supports both Winboard and auto232 and many other nice things (TASC board,
>>Mephisto board, DGT board, Fritz user interface, Chessbase analysis interface,
>>...)?
>>
>>Clearly different device level implementations of the same basic concept are
>>needed by people with mutually exclusive desires, so to solve this once and for
>>all a dynamical library would be perfect.
>>
>>Of course there is some work involved, but if you read this thread carefully you
>>will probably agree that again going straight for some kind of device driver
>>(Winboard protocol) is not the best possible solution.
>>
>>
>>
>>Moritz
>
>Maybe I wasn't very clear.  The protocol I have been discussing is *only*
>the protocol between the engine and the interface program.  To make sure that
>it includes enough richness in the command set to be able to accomplish the
>goal of playing completely legal chess, and being able to find out basic info
>about the opponent and clock and so forth.
>
>There is *nothing* here that would preclude making the "interface" program
>compatible with older auto232 programs, and I even mentioned that as one of
>the ideas.  There would be some "loss" of course, because the old auto232
>protocol doesn't include all the features I think we need.  But there should
>certainly be an "auto232" option in the interface program so that it can talk
>to an auto232 program over a serial connection.
>
>Definitely planned...
>
>I would first like to get a concise set of commands that everyone agrees is
>enough to play chess in any way we might like, from an electronic tournament
>director, to play on the chess servers, to whatever...

I have not much time to elaborate now, but you might take a look at what I have
done in TCB's chesslib for the CEngine class. This is exactly what is needed to
define a chess engine functionality. I think chesslib could be helpful for your
project since it would be possible to subclass CEngine to make a communication
with a console i/o-based protocol. This would instantly give to any chess engine
that conforms with this protocol all what I have programmed for TCB : windows
auto232 interface, winboard interface, cpip interface, windows gui, epd testing,
logging, etc.

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.