Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Fruit-IsiChess

Author: Dann Corbit

Date: 12:22:58 08/15/05

Go up one level in this thread


On August 15, 2005 at 14:59:44, Uri Blass wrote:

>On August 15, 2005 at 14:46:16, Cesar Contreras wrote:
>
>>
>>>I do not think that doing changes in the last days before the tournament is a
>>>mistake if you check the code for bugs and fabien does it.
>>
>>
>>It's easy to broke things in a complex system like a chess program and it's very
>>dificult to discover bugs, they are likely not to be discovered with just a few
>>days of testing.
>>
>>http://www.ece.cmu.edu/~koopman/des_s99/sw_reliability/#bathtub
>>
>>It's my believe that days before an important event it's better to focus on
>>adjusting parameters, book preparation and hardware preparation than changing
>>code.
>
>It is my believe that Fabien is a good programmer so he already has good tools
>to detect bugs and changes in the evaluation like adding contemt factor in the
>last day are not a big risk for him.

I don't care how good you are.
It's definitely a very bad idea.
In fact, it is a very bad idea to use untested changes in any production arena
at any time -- even if the changes are ten years old.

Most chess programmers are already very good programmers.  But the best
programmer in the world will make mistakes.  And if there is no time to test the
changes thoroughly, it is better not to try them.





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.