Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What is Botvinnik's legacy to computer chess?

Author: Christophe Theron

Date: 10:26:45 02/22/00

Go up one level in this thread


On February 22, 2000 at 11:03:00, Amir Ban wrote:

>On February 21, 2000 at 21:41:21, Christophe Theron wrote:
>
>>On February 20, 2000 at 17:22:12, Amir Ban wrote:
>>
>>>On February 20, 2000 at 14:49:58, Christophe Theron wrote:
>>>
>>>>On February 20, 2000 at 10:01:46, blass uri wrote:
>>>>
>>>>>On February 20, 2000 at 02:35:02, Bruce Moreland wrote:
>>>>>
>>>>>>On February 20, 2000 at 02:25:32, Eelco de Groot wrote:
>>>>>>
>>>>>>>Botvinnik worked for many years on his program Pioneer but had very poor
>>>>>>>hardware available to him in the USSR. It could solve some very difficult
>>>>>>>positions from Botwinnik's games but never reached the stage where it could play
>>>>>>>whole games as far as I know.
>>>>>>
>>>>>>The essence of intelligence is generalization, and the ability to generalize,
>>>>>>however poorly, is built into any chess program very early on.  Anyone can
>>>>>>create a program in under 24 hours that plays a complete game.
>>>>>
>>>>>
>>>>>I do not think that anyone can create a program in under 24 hours that plays a
>>>>>complete game of chess even if the task is only to choose a random move.
>>>>
>>>>
>>>>In 1987 I write the first version of my chess program for PC in one weekend. It
>>>>began to play games only a few hours after I started to work.
>>>>
>>>>Of course I already had the experience of writting a chess program, but this one
>>>>was completely different from the one I had written before. Not the same
>>>>computer, not the same programming language, not the same basic data structures,
>>>>everything was different.
>>>>
>>>>I think any experienced programmer, or even student, can create a chess program
>>>>in a very short time, and a program that can play reasonnable beginner's moves.
>>>>
>>>>It actually happens all the time. Creating a chess program is a rather common
>>>>project in the universities. A smart student can read some thesis about the
>>>>subject and quickly write his own program.
>>>>
>>>>
>>>>
>>>>>Maybe you are right about professional programmers but
>>>>>there are many people who do not know to create computer programs and many
>>>>>people are going to fail in the task of creating a chess program that play chess
>>>>>in under 24 hours even if they know something about programs but did only some
>>>>>simple programs of not more than some hundreds of lines.
>>>>
>>>>
>>>>Probably it's a difficult task for many people, but still it's doable and has
>>>>been done already by non-professionals.
>>>>
>>>>
>>>>
>>>>    Christophe
>>>
>>>
>>>You guys must be terribly bright. Well, actually, of course you are, but I have
>>>no idea what you are talking about.
>>>
>>>Writing a chess program from scratch in 24 hours or even a week doesn't make the
>>>slightest sense to me.
>>>
>>>I think if I lost all my sources and had to recode my program based on memory it
>>>would probably take me about 2 weeks to come up with a simplified but working
>>>version.
>>>
>>>I would need to build the basic data structures, code the move generation
>>>functions, code makemove and unmakemove, try to remember how this alphabeta
>>>worked and where I need to change the signs, work on the quiescence search,
>>>patch some simplified evaluation function (no hope of remembering even 10% of
>>>the real thing) and think out its internal data structures. Then work on game
>>>control structures, identify the terminal positions, do input & move parsing,
>>>and do some move and board display. Did I forget anything essential ? Probably,
>>>but I'll find this later when I start compiling and debugging, which will
>>>probably take a long time because I coded in a hurry.
>>>
>>>I expect the result to look quite unprofessional and to play rather weakly after
>>>only two weeks. If you want some fancy features like a working transposition
>>>table and an opening book, you'll have to give me an extra week.
>>>
>>>And all this is just to recode something that I already have and know well. If
>>>someone is merely a bright programmer but has to think out all these issues as
>>>he goes, how much do you expect this to take ? I think writing a chess program
>>>from scratch is certainly more than a student semester project, which takes a
>>>semester. I manage programmers on a daily basis, and I need to have a feeling
>>>for how long tasks will take them. I would not assign even my best programmer to
>>>write a working chess program in less than 3 months, and even that seems a bit
>>>ambitious.
>>>
>>>What am I missing ?
>>>
>>>Amir
>>
>>
>>What a lousy programmer you are! Let me tell you that you'd better speed up a
>>little bit if you want to have the slightest chance to enter the SSDF list. :)
>>:) :) :)
>>
>
>I understood that to improve is useless because my last name already
>disqualified me.
>
>:)


Don't be too unhappy with your name. When I was young I had a very serious
problem all the time: in arcade games, when you make a new high score, you have
to sign with 3 letters. It has always been a headache for me. It has never been
for you.

Life is so unfair! :)




>>More seriously, the discussion is about Botvinnik. In 20 years they did not
>>manage to produce a program that played chess, even poorly, when one can write a
>>poor chess program in 24 hours.
>>
>>That's all.
>>
>
>Botvinnik is not the first researcher to start an academic project without a
>clear idea of where he is going, with the expectation that hands on work would
>point out the direction. He's also not the first to discover that sometimes a
>vague idea just remains vague and nothing comes out of it. This happens in AI
>research all the time.
>
>I don't believe that writing a chess playing program was his obstacle. It's just
>that he had nothing to show and so no real reason to write it.


I agree. If he had the slightest thing to show, I think he would just have asked
a programmer to write a conventional engine, and added on top of this his
"expert human-like" problem solver. When the "expert" finds nothing special to
play, the conventional engine takes controls and plays.

This would have produced an average program with the ability from time to time
to play a flashing move. Always good to get publicity and credits in order to
continue the research work...

But obviously, he couldn't even do that. So this research was either very poorly
managed, or complete rubbish.



    Christophe



This page took 0.01 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.