Author: Andreas Herrmann
Date: 18:43:44 09/23/02
Go up one level in this thread
On September 23, 2002 at 20:49:25, Uri Blass wrote: >On September 23, 2002 at 20:04:13, Thomas Mayer wrote: > >>Hi Daniel, >> >>> Having said that, I'm also sceptic so far. But as far as I can tell, the only >>> thing which really speaks against Ruffian being original work is the fact >>> that.. well.. it doesn't happen every day that a very strong engine appears >>> and most of CCC-readers didn't hear of it before. That surely says something, >>> but it's very well possible that Ruffian _is_ the exception here. >> >>well, that's nothing against you, but to be honest: I absolutely do not >>understand the discussion about Ruffian the last view days... Instead to be >>happy that there is something new and something strong, it seems that most of us >>here are searching for a way to claim or proof that it is a clone... >> >>After having played some games with the engine I am absolutely sure that it is >>no clone. Also it does not play from outer space, but Ruffian is really really >>strong. In two little privat tests against Fritz 7 (6,5 - 3,5 in favor of Fritz) >>and Hiarcs 7.32 (7 - 3 in favor of Ruffian) it shows that it can bite. >> >>Besides: "most of CCC-readers didn't hear of it before" - well, might be a good >>idea for all of you to look a little bit more around where the heart of amateur >>computer chess beats, like FICS and ICC... I don't know when it was that I see >>Ruffian the first time there, but it is not as new as some think... And it is >>not only myself who has seen that this program IS strong... (e.g. I remember >>that the programmer of Tao remarked some weeks ago that it is very strong) >> >>And: why do you guys think that it is only for the pro's possible to write such >>strong engines - who knows, maybe Per-Ola has found something new... A couple of >>years ago nullmove was something new and it was a big step forward for computer >>chess... I am absolutely sure that there are still many things to discover in >>computer chess, maybe not as revolutionary then null move but they are there... > > >For me the problem is not having ideas but implementing knowledge of other(I >believe that if I understand how to implement knowledge of other people it will >also be more easy for me to implement my ideas). > >The next thing to do for me is implementing more winboard commands >and implementing hash tables in a better way. > >I wonder how did you learn the knowledge of other people about these things. > >Did you read source code of other programs? > >I even have problems in implementing non chess ideas that are only about reading >the winboard commands when I analyze. Hi Uri, i don't know what you are programming in your job. But writing interfaces or such things should some of the easiest things for a software developer. More difficult for a normal software developer like me, are only the special algorytms for chess programming. 90% of the source code in a chess program is absolutely unuseful in every standard application development like database GUIs and such things. Look to the engine-inf.htm from Tim Mann there is the whole wb protocoll and hints how to program, very well described. I have need before about 2 years only a few hours to implement the wb-1 interface with the most importtant commands into Holmes including logging in a file. And i am, as a delphi developer, have had no source code where i could have a look into. (very bad english - i know) If you want to implement the best way, use a second thread for the communication, like Tim Mann describes in the engine-inf.htm. Read the input buffer until you get a linefeed (#10) or the buffer is empty and store it in a own stack variable. This function you let run in a loop. So now if the main thread (= the chess engine) has time to work off the next command then the main thread takes the oldest command from your internal stack. That's absolutely easy to implement and has nothing to do with special computer chess programming. If you want i can send you my delphi code, but i think it doesn't help you as a C programmer. have a nice day Andreas > >I wonder how other learned all these. >The proble with learning source code is that there are so many files that are >dependent on other files that I do not know where to begin. > >Inspite of that problem I understood tscp but it was not easy for me. > >I started a discussion about listening to winboard commands when the program >analyze(things that are done in analysis or pondering and I have problems there) >see http://www.f11.parsimony.net/forum16635/messages/35202.htm for the first >post > >Uri
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.