Author: Robert Hyatt
Date: 13:59:33 04/07/05
Go up one level in this thread
On April 07, 2005 at 13:40:23, Steven Edwards wrote: >On April 07, 2005 at 11:34:49, Marc Lacrosse wrote: > >>R Hyatt wrote : >>>that is illegal PGN. Unless the standard has changed, the last time I looked >>>promotions were indicated as dxc8=R, _not_ dxc8R > >>You are right, Robert! > >Indeed he is. > >>In PGN standard 1994.03.12 at "8.2.3.3: Basic SAN move construction", one can >>read : "Pawn promotions are denoted by the equal sign "=" immediately following >>the destination square with a promoted piece letter (indicating one of knight, >>bishop, rook, or queen) immediately following the equal sign. " > >This is really a SAN issue and not (directly) a PGN issue. I do recall a lot of >discussion on this and several alternatives, plus the regular voting. > >One alternative that I liked was to use a solidus (forward slash "/") as the >prefix character to the promoted piece, in part because it had a long history of >use from EDN. But it was rejected. > >Another idea was to use a pair of parentheses around the promoted piece as this >also had some in print heritage. It was rejected in part because some thought >it expanded the SAN string unnecessarily. > >Then there was the alternative of not having any special indication at all. But >this was disliked because PGN, with the goal of being easily readable by humans, >was thought by some to require a special indication that could be quickly >located with a fast text editor search. > >>But maybe you should consider accepting the faulty syntax without the "=" sign, >>because : >>- it's the way Chessbase exports games to PGN (...sighing...) > >They should fix this; it would take at most a minute or two of coding. > >>- it is not easy to correct this "by hand" > >Nowadays I suspect that very little PGN movetext is entered by hand; most comes >from the output of mouse/GUI enabled programs. > >>- other engines (Yace, Pharaon, ...) seem to accept the "relaxed" syntax > >So does Symbolic's toolkit PGN reader. (Many other typical deviations are also >processed correctly.) But this also slows down the reader, and that's not a >good idea if you're frequently importing large amounts of PGN data. Crafty's >PGN reader is quite fast, certainly faster than Symbolic's, in part because it >doesn't try too many attempts at correction. Would you really want to slow it >down? Crafty already handles some non-PGN compliant things. for example zero-zero for castling rather than oh-oh. or 1.e4 rather than 1. e4 (missing space). You would think that software vendors could get this right, but then ChessBase has modified the PGN standard anyway with their new annotation tags and crap, so they will probably never plan on being 100% compatible with others... I'll try to fix this, but it is a real pain to waste my time fixing things that are caused by someone else's stupidity. :)
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.