Author: Miguel A. Ballicora
Date: 13:40:11 11/08/01
Go up one level in this thread
On November 08, 2001 at 15:40:18, Georg v. Zimmermann wrote: >On November 08, 2001 at 14:19:32, Ron Murawski wrote: > >> >>>I make most of the bug reports for Beowulf myself. However, because it's open >>>source then people like Ron Murawski, who want to write their own program, but >>>don't particularly feel like learning the coding required for the engine, can >>>just get straight down to writing the eval() function. After all, that's the >>>most fun bit, right? >>> >> >>Colin, >> >>Without the Beowulf open source project, the Horizon chess engine would not >>exist. I am riding on your back and Dann Corbit's and I will be forever grateful >>that you have made the Beowulf codebase public. >> >>What is great about open source code is its publicness. It is allowing me to go >>inside a clearly-written, working chess engine and I'm slowly learning how it >>all works. It's allowed me to insert a replacement evaluator and call it "my" >>program. So, open source has been good for me. >> >>I have pointed out several important evaluation errors in Beowulf and passed >>them on to you. I've also done beta-testing and reported general problems as >>well. Also, I have made suggestions about the evaluation of passed pawns and you >>have coded my ideas. So, open source has been good for you as well. >> >>I suspect that the friendly rivalry of "Who's evaluator is better?" is spurring >>us both along. The latest Beowulf is light-years better than its previous >>version and I hope to do the same with my own evaluator. >> >>When no one is trying to make a profit and everyone remains friends, then open >>source can succeed extremely well. >> >>Best regards, >>Ron >> >>PS: You're right, Colin: Writing the eval *is* the most fun thing! > >That sounds great! When I started I used to think eval() was most important. >Then at some point I swicthed to the search, where then extensions were the top >exiting thing. Then I realized you can only get that far with that, and improved >the razoring and most important move order. Then I tried some performance >improvement, which was easy cause the code was very slow. Only getting so far >since I suck at programming. Now I'd need to spend time on a testing >environment... > >I wonder if anyone not starting from scratch has learned in a different order ? I started from scratch and I followed your order. When I decided I wanted to have an engine I knew some things I wanted to code in the eval. For that reason, a bitboard approach sound the most logical way to go and determined a lot of things etc. etc. To make a long story short, the eval was the part of the program I dedicated less time!!! Now I am working on that. Of course, it took me ~ 5 years to get a complete engine. Regards, Miguel > >Georg
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.