Author: Dann Corbit
Date: 16:51:05 11/13/02
Go up one level in this thread
On November 13, 2002 at 19:30:04, Bob Durrett wrote: >On November 13, 2002 at 18:31:59, Dann Corbit wrote: > ><snip, snip and snip> > >OK. Can we reach agreement that the process of producing a chess engine is more >like the way Movie was made? Most software tools are created incrementally. >First the program was created to do certain basic essential things. Then the >programmer "taught the engine" to do more. After that, then another "lesson." >This process continues until the program became/becomes very strong. Some engines are strong out of the gate. Some engines never gather any strength, even after years of effort. >Presumably, the algorithms implemented in the earliest versions of Movie were >relatively simple. Then, with each "lesson," more complexity was added to it's >algorithms. At each step, the new algorithmic content was coded and the coding >debugged. Perhaps most of what is published about chess engine algorithms was >not incorporated into the earlier versions but maybe more of it incorporated >later. There are many distinct ways to make a chess engine stronger. One way is to improve the evaluation function. Doing this does not change the algorithms, but it can change the way the algorithms behave. For instance, if an improved evaluation function improves move ordering or has some other side effect on the program's execution, the existing algorithms may run more efficiently. >This process likely included some originality and innovation in the development >of the program's algorithms. Conceivably, the algorithms used are completely >different from what has been published. "Only the programmer knows for sure." > >In the beginning, Movie may have performed because the program [NOT the >programmer] was relative immature. The same way a human child grows. I don't think that programs grow the way that people do. I do think that programs grow in whatever direction the programmer wants. He may explore new algorithms or new evaluation terms or many other avenues. >Incidentally, when I talk about program maturity, I am not talking about >programmer maturity. Those are two different topics. I would add that programmer maturity may or may not affect programming skill. >How about that? I don't understand the question.
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.