Anonymous user
Talk:Chess player: Difference between revisions
→Task size: The problem is that longer solutions hinder readability
(→Task size: High quality chess playing is probably out of RC's scope) |
(→Task size: The problem is that longer solutions hinder readability) |
||
Line 10:
I think this task is too large, having reviewed and written several small chess programs in various languages. How about some smaller pieces: alpha-beta search, move generation, board representation, xboard or UCI protocol handler, EPD parser. Note that there are entire wikis devoted to just this task (http://chessprogramming.wikispaces.com).
: I don't think so. The PicoLisp solution has 447 lines (without the white space). There are tasks in RosettaCode with larger solutions. I feel it would be difficult to separate the pieces, as they strongly depend on each other. --[[User:Abu|Abu]] 18:04, 24 April 2012 (UTC)
:: I think some languages have solutions (and output traces!) that are far too long; I like tasks to be such that at least the majority of solutions will be able to fit on a single screen, so you can see the whole of them at once. (Can't be done with some languages; for example, [[C]] is usually fairly verbose.) Whilst still being idiomatic; that's ''important''. That said, a player (even if it only chooses randomly from the legal moves) would be a reasonable start. That would then be a start of what any more sophisticated player would need. –[[User:Dkf|Donal Fellows]] 14:04, 25 April 2012 (UTC)
I would love to see folks attempt a modern chess engine in other languages though. My recent survey found that all the top engines were in C/C++ (rated 3000-3300), with the [[Delphi]] engine Booot at 2935, [[Java]] engine Cuckoo at 2675, and [[C sharp|C#]] engine Pupsi at 2610. Engines in other languages were only of amateur demo quality, not even worth mentioning. --[[User:IanOsgood|IanOsgood]] 17:21, 24 April 2012 (UTC)
| |||