ok I looked through this and also this (http://www.saremba.de/chessgml/why.htm) which is reference. What I find truely hilarious is that NEITHER of the authors seem to understand the actual usage of the pgn format. They are missing things in their specification like alternative lines and commentary (both very important for the game of chess). In other words you need to represent alternate moves and comments WITH the move thats being made, simply representing the header as xml doesn't gain a whole lot in maintaining states throughout the move list.

I personally find the fact that chessgml uses SAN very usefull ... now in order to be able to actually read the moves you need to know the rules of chess ...

example ... Nd3 you need to know to move the Knight to d3 (which means you need to find the Knight on the board, if there are two of them only one will be able to make the move so you need to determine the valid moves for a knight and find which knight might be able to move there)

It would seem to me that this "standard" is a bit too narrow minded, how many similar games are there to chess (Go, Tic Tac Toe, and Checkers to name 2) why not include the definition of the board, pieces, player counts, etc in the header and make a generic standard for a grid based board game ? You could then use simple origin-destination notation to represent the moves and completely remove any requirement of knowing how the pieces move or knowledge of the current board.

You could also easily do alternate lines and commentary by making move recursive

<Move src="a1" dest="c3"piece="a pieceidentifier">
<AlternateMoves>
<Move src="a1" dest="c2" piece="a pieceidentifier">
/* you could follow a complete alternate line through her */
</Move>
/* a piece being put onto the board for something like tic tac toe or go */
<Move src="" dest="c2"piece="a pieceidentifier">
</Move>
</AlternateMoves>
</Move>

Also there is little discussion on the use of such formats with computer engines ... the above format could also be used to represent an opening book for a computer engine (prior to being hashed) this would be very easy to turn into an ECO book of variations ...

just to be clear when I say a piece identifier I mean an identifier to a piece which was defined prior in the document (included)

I should also add that a Move would contain comments and possibly results / game events which could be defined by the game include.

heh too much posting as I correct details, it being that you are building a board, moves should be similar to coordinate notation where they are both numbers seperated (row and col) as opposed to requiring one axis to be letter and the other number.

Greg,

Thanks I guess someday I shall talk to you to understand Chess much better! I believe you should be a guru in chess to go to such extents to elaborate your thoughts :)

I have dabbled with chess a bit (mainly on the engine side) ... this article though gave me some good thoughts and I am currently working on a small implementation, thanks :)

The 9478 Laura Williams blog

The 5185 Kimberly Roberts blog

The 4254 Charles King blog

The 7834 Laura Williams blog

The 3047 Edward Hernandez blog

The 1528 Joseph Baker blog

The 8661 George Walker blog

Yo!
New here and first post.
I am boy, 21 years old from Ireland
I enjoyed here.

Botijos