Thanks Drew, for your links.

-------------------------------------------------------------------
Drew wrote:
-------------------------------------------------------------------
> Personally I favour a simple INI-file style i.e. key-value pairs and
> sections; terser even than PGN

Yes, this is another idea, and probably I will go this way. But
using XML syntax has the advantage that I can let the XML
parser doing the syntactical stuff, except reading the game
moves of course. So my first try is to use XML, I like to have
a format as simple as possible, and XML is providing this.

> In general XML is very verbose compared to something as terse as PGN.

1. I don't think so if I don't use <move...> markups.

2. It is planned to invent an additional binary format, based on
   the text format, with a very simple and fast en/decoding. I've
   already have a concept for this.

> Indeed what is your problem with PGN

See 
<http://talkchess.com/forum/viewtopic.php?topic_view=threads&p=562958&t=51666>
about the problems with PGN. And I have forgotten one more important
argument: support of multilingual comments, PGN doesn't know anything
about other languages than English.

Another problem with PGN is the complexity of parsing, have a look
into Scidb's implementation of PGN file parsing:
<http://sourceforge.net/p/scidb/code/HEAD/tree/trunk/src/db/db_pgn_reader.cpp>.
This is a catastrophe, PGN has grown uncontrolled because of
deficiencies. The new format will be very simple to parse.

> especially if it is only for your own purposes.

Probably some other applications will adopt the new format as
an additional option, I'm not the only one who needs a more
powerful format than PGN.

> http://www.cybercom.net/~zbrad/Chess/pgnxml/

This project is using the syntax
<move n="nn" state="xx" annotation="xx" promote="x" from="xx" to="xx" NAG="nn" 
display="xx">
for game moves. Although this is strong XML, it's too much, the
file will explode. Even zipped PGN files sometimes will be very
large. But the idea how to handle the tags seems to be a good model,
for example:
<White Elo="nn" USCF="nn" NA="xx" Country="xx" Type="xx" Title="xx">

> http://www.saremba.de/chessgml/

This project is abandoned, and the syntax will cause exploding file
sizes.

> http://xml.coverpages.org/chessML.html

This project is abandoned, and all links are broken on this page.

-------------------------------------------------------------------
Pietro wrote:
-------------------------------------------------------------------
> I had some experience with xml and java package for parsing it,
> but I do not understand why you need another (apart of .pgn. ,
> .sci., cbh, .si4) format.

This new format will serve two purposes:

1. Backup of single games in a human readable format. PGN is not
   supporting the features of Scidb.

2. Interchanging data between chess applications. Of course this is
   possible only if this format will be widespread accepted. PGN
   cannot be used for this purpose, no way to import an exported
   game from ChessBase via PGN, PGN doesn't know anything about
   multilingual comments and different comment types like
   post-comments and pre-comments.

Gregor

Reply via email to