http://ccif.sourceforge.netIt is still a draft, but I think it is very interesting, and possibly technically superior, but not sure if has better compression than scid.
Just realised that the author of the new format is actually the author of scidb [Gregor Cramer].
This gives the new format a lot more credibility.
It's XML-based though, thus it's not very human friendly, is it?
The primary intended use of a chess interchange format is for archiving games and interfacing with GUIs, not for humans to read the code.
Thus I think a focus on ease of parsing for GUIs, speed, compression and extensibility to other chess variants is the primary goal of C/CIF, and it excels at this.
If you look at it from lichess development perspective, it greatly simplifies a lot of the pgn parsing, and at the same time allows for the use of variants that were not possible before on the website due to the limitations of pgn.
XML is rubbish in many ways and is basically depreciated nowadays, in favor of JSON and YAML at least.
I would agree. JSON and YAML are much cleaner.
> If you look at it from lichess development perspective, it greatly simplifies a lot of the pgn parsing, and at the same time allows for the use of variants that were not possible before on the website due to the limitations of pgn.
Maybe but if this new format is not human-friendly, it will NOT replace PGN, so you'll always have to parse PGN anyway.
PGN isn't used on lichess backend. PGN is only used when allowing players to import/export their games - as an interface.