Author: David Rasmussen
Date: 11:50:41 02/01/03
Go up one level in this thread
On February 01, 2003 at 14:27:57, Russell Reagan wrote: > >I look at it from the other way around. If a person posts a FEN string, it can >be assumed that there have been no position repetitions, and then the fifty move >counter does have some meaning. What do you mean "there have been no position repetitions"? The important thing to know is not whether there have been repetitions, but what positions have been on the board before the position, but after the last irreversible move. If the halfmove clock is h (say, 27), then pick some random position with the last h positions, call this position P. The program given the FEN will not know that after reaching position P twice when searching this position, it is a draw. You can only do two things: Assume that the halfmove clock is always 0 in a FEN, no matter what it says, or assume that the h positions played before this one can't be repeated (you say that the last 27 positions have the hashkey 0, and the program will never see a repetition, because no position (hopefully) have a hashkey of 0). None of these are optimal, but since repetitions are almost always important in positions with a signinficant halfmove count, in fact scratch almost, repetitions are _always_ important in this case, I think the best thing to do is to ignore the halfmove clock. So I would rather say like this: FEN shouldn't have halfmove clock. If the halfmove clock is not 0, then post a PGN that starts from the last irreversible move, and then give the moves since that position. >Otherwise the person posting should have posted >the PGN, or is just trolling and trying to waste everyone's time. It is the >fault of that person for not providing the necessary information, not the fault >of the FEN standard. I disagree. It is a very artificial thing to have halfmove clock in the FEN. It is _never_ useful. Show me a FEN where it is, and I can always say "some positions must have been encountered before this, that can be repeated, tell me what they are". >No need to change the FEN standard. That is what PGN is >for. It seems like you are trying to change FEN into PGN. Just use PGN if you >have such a prblem with FEN. > PGN is good in this case, I am just saying that halfmove clock shouldn't be a part of FEN for that very reason, and the people should always set it to 0. >The FEN string tries to contain as much useful information as is reasonably >allowed to maintain a compact representation of a position. The game history >removes the compact aspect. > When is it useful to have halfmove count in a FEN? >I'm still not sure what point you're trying to make or what result you are >hoping to acheive. FEN serves it's purpose well, and for the instances where FEN >is insufficient, there is PGN. I am just discussing FEN, that's all. First step was to post my point. If everybody agreed, then _maybe_ we could think about changing things (A chess application XML specification would be a cool thing anyway, but that's another discussion). Now, people do not seem to agree, so I am still explaining my point. Seems reasonable, doesn't it? /David
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.