A game with free code, but non-free art & music - why not in the repos?
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
I noticed that "Frogatto"( a 2d jump'n'run game) is part of the ubuntu and debian repos, but not of trisquel.
The developers write on their website (http://www.frogatto.com/), that the code is released under the GPL, but not the rest of the game (art & music).
Because of this, the main package of frogatto is in debians "contrib" repository, because it depends on "frogatto-data", which is non-free.
I listened to many, many talks of rms and he mentioned very often (and I agreed very often) that the art of a game is a different thing - it's not necessary that this part of a game is free as well!
So in my opinion, the developers of frogatto do just the right thing.
It would be sad if the game isn't part of the trisquel repos because of ethical issues.
Or are there some other reasons?
I should underline that I do not complain about the game not being in the repos - I just want to make sure that this is not because of ethical issues.
I initially believed the reason is purely technical: since it seems to be in ubuntu "multiverse" repository (implying non-free) it won't get included.
However, when looking at the copyright file of the frogatto package [1] I don't understand why it's in multiverse. Perhaps they chuck stuff from debian's contrib there?
Looks like frogatto-data has a crazy non-commercial and no-derivatives license conflicting the earlier license... [2] It's hard to understand what the heck they actually mean. Why can't people just use standard licenses?
So in the end I'm not sure at all how frogatto and frogatto-data are licensed. Lovely!
I just read a bit here: http://www.frogatto.com/developer
and found this passage:
"The license on the code is the GPL, but we’re doing a Qt-style dual-licensing model, so we insist that if you contribute something, you transfer the copyright to us. We of course immediately re-release it in our source tree as GPL, as you would want."
I don't quite understand what they mean. Why should anyone transfer the copyright to them?
Perhaps this weird passage is the issue why frogatto is not in trisquel.
Copyright assignments to companies are often bad, enable harmful
business models and discourage merging changes from other developers.
This is not a reason to exclude a package from Trisquel (although many,
like some Oracle-copyrighted programs, have now better forks without
copyright assignment).
The problem with Frogatto is that its data license forbids selling
(e.g. of DVDs with our packages if it was included in the distro). We
accept nonfunctional data (e.g. game art) if it allows possibly
commercial distribution of unmodified copies.
"non-functional" data that is not free is accepted by GNU Guidelines for free software distributions[1]. But, the license must give you permission to copy and redistribute, both for commercial and non-commercial purposes. It seems frogatto doesn't allow that.
[1]: http://www.gnu.org/distros/free-system-distribution-guidelines.en.html#non-functional-data
I would add that only the C++ code is free. The non-free data includes Frogatto Formula Language code.
Can you specify this a little bit more? Which kind of language code is implemented and how did you find out?
Good question. Here's a link to the code on Github:
https://github.com/frogatto/frogatto
The LICENSE file says that only the C++ code in 'src' is under the GPL. The 'data' directory contains FFL code.
I found out way back when the game was announced on the Wesnoth forums before the 1.0 release.
I took a look, and although there is some basic math and some sort of function calling, I'm not so sure that those are software. For the most part, they're just data files, with a format similar to JSON. Any actual function must come from the main program.
Unless there's something I'm missing?
IMO, math and function calls definately count as software code.
Besides, the FSF guidelines require that data be licensed to allow commercial use, and the Frogatto license forbids commercial use.
> IMO, math and function calls definately count as software code.
Why would you say that? You wouldn't say that the math problems on a homework sheet count as software. Why should a mathematical expression be treated differently simply because software uses it?
Function calls only do what their definitions tell them to do. The function calls themselves only give information: the name of the function, and a list of arguments. I don't see how simply listing a function call in a file counts as software.
From an abstract point of view (i.e., forgetting about the hardware limitations), programming is like proving a theorem in a constructive way (i.e., constructing an actual solution to the problem). That actually is the fundamental reason why software, like mathematics, should not be patentable.
The abstract problem is instantiated in a way that makes it solve *your* particular problem. It is crunching *your* numbers. You deserve control over it. All four freedoms.
As for the function declarations, I agree. An API should never be copyrightable. At least for interoperability purposes.
- Vous devez vous identifier ou créer un compte pour écrire des commentaires