Author: Uri Blass
Date: 14:04:08 11/16/02
Go up one level in this thread
On November 16, 2002 at 16:28:14, Anthony Cozzie wrote: >On November 16, 2002 at 13:16:33, Bob Durrett wrote: > >>On November 16, 2002 at 12:29:35, Peter Skinner wrote: >> >>>On November 16, 2002 at 11:45:27, Bob Durrett wrote: >>> >>>> >>>>I'm Confused. C, C++, C#. Which to Use? >>>> >>>>I have an old Visual C++ compiler which I purchased many ago. But bulletins >>>>posted here seem to favor C, as opposed to C++ as if C is better than C++ for >>>>chess engine programming. >>>> >>>>Still trying to find the best C compiler to use for chess engine design on my >>>>Compaq 1 GHz 1GB RAM single-processor PC. >>>> >>>>Tried to download one but got error message saying NO! >>>> >>>>Yahoo! seems to find C++ compilers but not C compilers. >>>> >>>>Sorry for sounding like a child. : ( >>>> >>>>Bob D. >>> >>>About 2 to 3 years ago I wanted to write a program as well. Since I had no clue >>>where to start I based my engine off the GNUChess 4 source. I was learning C the >>>same time so it was nice to have a code to start off of. >>> >>>After 3 years I would never try to start from scratch. There are two main >>>reasons: >>> >>>1. I am not a very skilled programmer. I can implement things, but I ask a >>>million questions like yourself, and there is nothing wrong with that. People >>>here are great with answering questions. The answers I have recieved through >>>this board and the Crafty mailing list have helped immensely. >>> >>>2. Time. I don't have much time to program. Starting an engine from scratch >>>takes an incrdible amount of time. I would suggest that you look at some sources >>>to understand the different structures that Gerbil, Crafty, GnuChess anything >>>that has published source code. That helped me quite a bit. Those authors who >>>_have_ put the time in to start from scratch give the code out for people to >>>learn from and I am grateful. >>> >>>Personally if you want to write an engine, I would start in C. Most compiler >>>that are for C++ are C complilers anyway. I prefer the Visual Studio from MS as >>>the MSDN library has some great code examples to look at. >>> >>>That's my two cents. >>> >>>Peter. >> >>Thanks. The Microsoft Visual Studio's System Requirements seem to exclude >>Windows 98, which is what I have. Guess I'll have to buy a new $5000 chess >>computer. Convincing my wife is going to be a problem. I wonder if Visual >>Studio will run on a dual or quad processor machine? Anyway, I'm just joking. >>Maybe I don't need Visual Studio today after all. A pity, though. But . . . so >>be it. >> >>Bob D. > >My experience: I became interested in chess at about 11th grade, which was the >same time I learned to program in C. I of course tried to write a program to >play chess. Result: zip. I tried again during my freshman year of college. >Result: a move generator capable of generating 10,000 moves/second. The first move generator of me was even slower but the main reason is that I decided to do legal move generator and in the first exe file it did not do incremental calculation of the attack tables. I knew the problem but writing legal move generator without bugs is an hard task so I did not care about speed but only about lack of bugs(I use the perft function to check for bugs) and only later rewrote functions to do things faster) Today movei's speed in nodes per second is similiar to tscp. I could work more on improving it's speed but I decided that it is enough and I can start a chess program. Went back >to college. After my sophmore year, I tried again. This time I actually read >up on what other people, and I wrote an engine that actually did play chess >(rating about 1600). I actually precomputed stuff for the move generation, and >did some other things. Unfortunately I had massive transposition table and >quiescence search bugs. Finally, after my junior year I started from scratch >(again, sigh) using bitboards. I wrote a static exchange evaluator, which did >wonders for my Q search. Result: zappa. Even during zappa's creation, I have >rewritten parts from scratch (3 tries at the move generator, for example). > >The point is, writing a chess engine is *hard*. This is your opinion. I also find it as not easy but it does not seem to be the opinion of some other programmers(Tom Keriggen the programmer of tscp used only few hours to write his simple chess program) Uri
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.