Author: Peter Klausler
Date: 08:32:17 06/10/98
Go up one level in this thread
On June 10, 1998 at 03:06:27, Steven J. Edwards wrote: >On June 09, 1998 at 19:51:03, Robert Hyatt wrote: > >>if we are going to nitpick the standard, here is one thing that is >>really ugly: when you output a variation with black to move, it >>apparently should be done like this: 15... Nf6 Nf3 d5 etc. That >>"15..." is totally illogical. for normal moves we see 15. e4 h6 16. >>Nf6 and so forth.. so that the move number is one "token", then it >>must be followed by a white move, and a black move. Using "15..." >>violates that and is more difficult to parse to boot, since most >>scanf() input readers will suck in 15... as one string... >> >>why not use the more common 15. ... Nf6 format, where a separate >>token (...) indicates that the white move is omitted. This seems more >>common in published analysis and it is much more compatible with normal >>parsing algorithms... > >I have strong agreement with this as it does make things more >consistent. Also, there is a precedent in ANSI C with the same ... >elipsis token (for variable count formal parameter list declarations). > >The print literature is divided on this. (As it is on most things.) >Personally, I would like to not have any period immediately following >the fullmove number, but the voting back in 1992 was well in favor of >making move numbers look like decimals. > >-- Steven (sje@mv.mv.com) The practical result, unfortunately and perhaps unavoidably, has been that real parsing routines have to accomodate all of the above, simple to perform an (optional) check that the move number is not incorrect. The fullmove number and ellipsis carry no other semantic value. The more that I think about this, the less hopeful I am that any textual representation is going to be able to apply rigorous enforcement of rules on input without being undermined by another implementation. *Any textual representation is defined in a de facto way by its most permissive implementation.* If program X can read a file, then any program that can't will generate complaints. Perhaps the needs of interchange would be addressed better by defining a standard binary format, or by publishing an open specification of an existing proprietary binary format. What do you all think? -Peter Klausler, the CDB guy
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.