Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Is SMIRF compatible with Arena?

Author: Reinhard Scharnagl

Date: 22:05:33 07/26/05

Go up one level in this thread


On July 26, 2005 at 18:11:44, Uri Blass wrote:

>On July 26, 2005 at 15:23:17, Reinhard Scharnagl wrote:
>
>>On July 26, 2005 at 15:13:27, Jorge Pichard wrote:
>>
>>>I would like to test SMIRF against other programs like Glaurung-FRC 0.2.4 and
>>>Pharaon-FRC 3.3 under Arena, is that possible?
>>>
>>>Jorge

>>As repeatedly has been explained: SMIRF is an 8x8 AND 10x8 enabled engine. It is
>>using only ONE engine NOT TWO. Thus it could not use any traditional protocol.
>
>I still do not understand why it cannot use any traditional protocol.
>Arena does not support 10*8 but I think that a program can assume that it play
>8*8 chess if it does not get different information.
>
>You do not have to support winboard or UCI but I see no reason that it is
>impossible to do it with one engine that can play both 8*8 and 10*8 chess
>
>Uri

Hi Uri,

the main idea of SMIRF (whichs development now is pausing for a while) is to
SIMULTANEOUSLY support a lot of 10x8 AND 8x8 chess variants. To support a
protocol from the beginning like UCI would implicate to hide those abilities
from the user. Then it would have made more sense to develop an engine like
usual for UCI. Why should I write a new type of engine and then disable its
special abilities?

Nevertheless I had planned to use a UCI based protocol later. But because UCI
for long did not even support Chess960 and Winboard has been contradicting
concerning it, I decided to wait until an appropriate specification would have
been made. But when now SMK has published a UCI protocol extension according to
an unnatural view of Chess960 to be an incompatible variant instead of a
compatible superset to chess the idea to somehow use a UCI like protocol in
parallel has been canceled.

SMIRF though using TMCI is showing by doing that it is possible to support a lot
of variants WITHOUT SWITCHING the engine's mode simply by communicating with
compatible FEN strings and encoded moves. This idea also could have been
applicated to a UCI protocol extension. But actually that switches an engine
into an FRC mode forcing the engine to different encoding of moves and FEN
strings, e.g. generating different FENs for identic positions (where castling
rights are based on traditional K+R placements only).

For me it is unbelievable that engine programmers now are jumping onto that
train of incompatibilities. Having planned to integrate an 8x8 UCI support in
SMIRF somtimes later now is no more an option at all, reflecting SMKs attack to
the Chess960 compatibility of existing solutions, online servers, game
databases, PGN viewers, existing lists of starting positions etc..

Reinhard.




This page took 0.01 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.