rhollay wrote:But chess is a competition itself! White against Black.
Getting involved in computer chess, you cannot avoid competition.
For me, chess and also computer chess is a hobby, a _game_. Without competition between engine writers it would be just booooooring.
My point was that GPL-ed (or other open sourced) engines just hurt the competitive soul of this hobby, because reveal "secrets".
Exactly these "secrets" can make the competitive edge of the game.
Of course there is an element of competition. But why should there be secrets? The competition is not about who has the most secrets, but who has the best ideas. Fruit was a very revolutionary program, yet there were arguably no secrets at all in it.
To me it is like this: If I were to play against Fruit in a tournament, then it would be matching my creation--all of the algorithms I've used and how I implemented them and how they all fit together as a whole--against Fabien's. Both of ours' ideas are out in the open, so we know what the other did. The question becomes who came up with better algorithms. Because with open source programs, the incentive is not to copy other programs, but to come up with something new. For this reason, I would have absolutely no interest in playing against a Fruit derivative in a tournament. Because whoever is entering it, even if their program is open source, and even it is 500 points stronger, is simply cheating. For a closed source program, the incentive is just to make the thing stronger, not to come up with something new. Whether a person gets a stronger engine by borrowing from an open source program or by coming up with completely new algorithms does not matter, when this distinction is of utmost importance in open source. Consider Sloppy. It does not copy one line of code from another engine (I think), but the evaluation algorithm is taken from Fruit. It is strong, but it was mostly shunned by the community. I myself am a bit on the fence about it: because evaluation is my engine's weakpoint, I know that I could vastly improve my engine by doing the same thing. But it would be an absolute bore to me. I'd much rather come up with something original, so I choose to have a "crappy" engine. So it's quite clear: in the open source chess community, there is a big emphasis on originality, arguably more than in closed source. Of course, there are completely valid reasons for keeping source closed. But increasing competition is not one of them.
Right now, for whatever reason, Rybka is so far ahead of the other engines that most other programmers seem to have given up. After R3 has been released, I imagine the sales of other programs have gone down quite a bit, and many other programmers are reluctant to release anything because it just won't stack up. In short: there's no competition.
At least I can be glad that all the work I do makes computer chess boring to some people...
I sincerely hope that most people don't share your opinion. Luckily I don't think that they do.
