Saturday 20 February 2021

The joy (and pain) of Open Source Software

 Over the last 6 months I have been filling a kind of tech support/software development role for some online chess competitions (FIDE Online Olympiad, 4NCL Online, FIDE Corporate World Championships). One of the pieces of software I wrote was a tool to take a pgn generated by Swiss Manager and download the game scores from lichess.org and chess.com. This saved a lot of time for organisers, as it meant that there was no need to manually cut and paste games from server to pgn file.

I've happily shared it with organisers, and in some cases they have made their own modifications to suit the specific needs of the event (eg the 4NCL Online Junior involves 2 games between players, as White and Black). Through this process a number of issues have been identified and solutions found. To me, this is how Open Source software should work.

However, sometimes there is a downside to this. If you use software developed by others, then suitable attribution should be given. On occasion this has not happened, with 'authors' claiming credit for the work of others. One extreme case I was aware of involved passing off a piece of open source software as a new 'proprietary' product and selling it for $50,000 a copy. 

Recently claims have been made about 'Fat Fritz' and 'Fat Fritz 2' which have been released by Chessbase. Both products were based on previous open source projects (Leela Zero and Stockfish), with minimal changes to the code, but with different neural net data. While saying that a a new version is better than an old one is not forbidden, or even selling open source software for a profit (as long as the details of its origin are made freely available), erasing the names of earlier authors is. If you want to find out more, then here is the source article.


No comments: