Author: Dann Corbit
Date: 13:18:45 05/29/01
Go up one level in this thread
On May 28, 2001 at 19:06:01, Bruce Moreland wrote:
>On May 28, 2001 at 13:58:05, Miguel A. Ballicora wrote:
>
>>I am not sure I understand the basis of this discussion. I am sure
>>that I am missing something since I do not use EGTBs yet.
>>I'd like to understand because one day I might want to implement it.
>>What is the difference with Edwards and Thompson's tablebases?
>>I smelled that the complain is that everybody is using the _same_ code
>>(cut and paste) to extract information from Nalimov's but in Thompson's and
>>Edward's everyone used different code? (gotta be same algorithm though).
>>Is that so?
>>What did you write new? you still use the Nalimov's table?
>
>I wrote my own stuff. It was very hard to do. I haven't pushed this issue much
>because I don't want people to think that I'm trying to argue that I should be
>allowed to use endgame tables and others shouldn't. I wrote my own in part
>because I don't want an engine co-author in any form.
>
>I think the issue is worth considering because here you have a bunch of programs
>using the same source code to evaluate positions. This is not a good precedent.
I don't think it is good or bad. It is a choice, and probably a weak one. The
programs that innovate are the ones that advance. Look at where the clones are
on the strength chart right now. They slowly slip away into oblivion. That is
because someone who copies without knowing end-to-end mechanics of the system
will not be able to improve the system. At some point, they will lose.
If someone has a good idea on eval (let's say a check for hanging pieces) and
someone else implements that after the author describes it, can you really
imagine that is a bad thing?
I think everyone agrees that wholesale copying is bad. If you claim the idea as
your own, it is (in fact) criminal. But sharing of ideas and using code with
permission is not a bad thing at all.
You have put far more effort into your chess program than probably most people
have. This is especially true if you have implented every algorithm without
even looking at someone else's code {though surely you have read technical
documents explaining the ideas}. Perhaps you have an emotional stake in making
people suffer the same as you. Now, this suffering is to *good* purpose without
any question. Someone who goes through the same stages will definitely benefit.
On a couple different occasions, I received permission to modify someone else's
engine as a starting point. On both occasions, I later decided it was not a
good idea. I am working on my own design, which is comically different from any
that I know of. Maybe something good will come of it and maybe not, as far as
program strength. But I do know that this approach is far better for learning.
At any rate, I think if someone else wants to save some time by studying the
work of others, there is no evil intent there. And if they want to use someone
else's work *with permission and credit* I have no problem with that either.
>But this is a minor issue compared to shared opening books, since the opening
>book is important in *every* game.
More important than most people think, unless I am very much mistaken.
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.