Author: Bruce Moreland
Date: 09:20:25 06/15/98
Go up one level in this thread
On June 15, 1998 at 01:47:52, Steven J. Edwards wrote: >Fellow Programmers: > >In working with Crafty's opening book PGN file wall.pgn, I noticed an >error in SAN in one of the games. (There may be more than one; this is >the first.) It shows up at line 3,864,317 in column 47; it is a pawn >capture "exf1" but has no promotion indication. > >Have other PGN readers choked on this? If not, then they should have. >It is incorrect to assume a promotion to a queen. > >Anyway, having nearly four million lines of good PGN before the first >problem is still pretty good. Why should they choke on this? I thought the idea is that an importer should be flexible about input format? If it's not supposed to understand that 99% of the time it sees "exf1" it means "exf1=Q", and handle it appropriately, then what else is it not supposed to be able to eat? Should "1.e4" raise a syntax error? How about "1. nf3"? How about "1. e4 e5 2. Nc3 Nc6 3. N1e2" ? A problem I had trying to read this file is the existence of the "LongECO" tag, which is unspecified. I don't understand why the specification allows unspecified tags, which would seem to prevent you from ever finding typos in the tag fields, when someone enters "Evnet" instead of "Event". bruce
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.