Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The death of computerchess.

Author: Dann Corbit

Date: 18:59:50 12/20/01

Go up one level in this thread


On December 20, 2001 at 21:53:53, Christophe Theron wrote:
[snip]
>You just have to read the questions asked here by people writing new engines.
>
>It's all about very classic stuffs. Alpha beta, move ordering, hash tables...
>
>Nothing new.
>
>All these questions can be asked by: "look into Crafty source code, copy,
>paste".

This is commical.  Can you show any example where anyone has done this at any
time?  Maybe it will work for some tiny algorithm snippet, but not for any major
function.

If you are going to take some idea, it will have to work in your framework.
Crafty is a giant, interconnected machine with dozens of global variables.

I think you greatly overestimate the actual ease of an operation like that.

When I want to try something new, I *do* look to see what other people did.  I
might trace their code in a debugger if I don't understand it.  And I will look
at as many different implementations as I have.  Then I will read a paper on it
(yes, I do it upside down).  Then I will trace it again.  Now I know how it
works.  This is very much easier to do than writing it from scratch.  I will
always see things that I would do differently.  I will always see things that I
don't quite know why they did it a certain way.  In any case, I doubt very much
if anyone (outside of a pure cheater who just copies the whole thing and then
tweaks) will really try to do this.

It sounds simple to do.  In fact, I think it would be a great deal harder than
doing it yourself.




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.