Author: Steven J. Edwards
Date: 11:00:48 06/02/98
Go up one level in this thread
On June 02, 1998 at 11:00:14, Andreas Stabel wrote: >I just want to clarify a couple of things conserning todays PGN standard >that perhaps also may affect PGN 98. > >The first is the use of symbols to represent that something is not >known. >The PGN standard specifies several different symbols depending on the >place >you need an unknown symbol and I think that is bad. When I parse PGN >files >I see the symbols '?', '*', '-', ' ' and even some others to signify >that something like the result is not known. There are several places >where an >unknown value is needed like the result, date, site, event and even name >tags, and if my interpretation of the standard is correct you may use at >least three of the above symbols to show that something is unknown in >todays standard. > This is recognized. There is no problem with the game termination marker for unknown which is always "*" (without the quotes). PGN tag value for unknown and default values needs work. For PGN tags, the standard could be (and will be) more clear. >The second is from a discussion I've had recently with Mr. Hyatt about >how >the black move number should be written. Crafty outputs 15. ... Nf3 and >my >opinion is that 15... Nf3 is the correct way. The point is that both >Mr. Hyatt and me think that we have read the standard carefully to come >up >with our view so the standard is perhaps not clear enough on this point. Yes, there is an ambiguity here. The form "15..." is recommended. >Finally I want to say that whatever the new PGN standard is decided to >be >it has to be strictly and formally defined with no ambiguity and that as >far as possible it should be backwards compatible. I must say I became a >bit worried when you said that there will be no difference between the >import end export format. Perhaps you could expand on this. The motivation for having an export format was twofold: 1) A rigorous definition of the minimum PGN an importer would have to accept. 2) A byte by byte reproducible string that would be the same for the same game for any exporter running on any machine (subject to newline representation). The problem is that the specification was diluted by appearing both as to what could be read and what should be written. Import/export distinctions are eliminated in the upcoming standard; the specification will be stronger and less connected to external representation. -- Steven (sje@mv.mv.com)
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.