FreeGameDev

Moderated: Split from here: viewtopic.php?f=3&t=4782&p=49515#p49515


unfortunately SDL2 has changed from GPL to zlib, meaning that it is likely to aid the development of nonfree games -- that'll probably be distributed via a DRM-infested platform serviced through a proprietary program (i.e. Steam). this is a *big* mistake.

alexander {l Wrote}:unfortunately SDL2 has changed from GPL to zlib, meaning that it is likely to aid the development of nonfree games -- that'll probably be distributed via a DRM-infested platform serviced through a proprietary program (i.e. Steam). this is a *big* mistake.

SDL has never been licensed under the GPL, and SDL 1.2 was already used by several closed source titles, out of which some used DRM.

Back to the topic, though. The feature set of SDL 2 looks really promising, and there already seem to be a couple of Ubuntu PPAs that package it. I might give it a try at some point since I'm not particularly pleased with a certain other library that I have been using.

@amuzen Sorry, I meant LGPL. That licence choice was a smaller mistake. Changing to zlib is on the other hand a *big* mistake, and completely the wrong direction where user freedom is concerned.

alexander {l Wrote}:@amuzen Sorry, I meant LGPL. That licence choice was a smaller mistake. Changing to zlib is on the other hand a *big* mistake, and completely the wrong direction where user freedom is concerned.

Your worldview is impractical. The vast majority of games made for money (as in, games that pay bills) are not open source. By using a viral license, SDL would marginalize itself massively.

Inclusion in top titles will lead to contributions back to SDL (no developer would sanely want to maintain their own branch of SDL long term) which will benefit everybody. SDL supports Android and iOS and Mac OS X. These platforms can only be well supported if SDL is an industry toolkit - which the GPL would preclude.

I wish all software was open source. That's a fantasy though.

SDL being used in non-free applications is only a good thing, as it promotes SDL, thus, aiding in both platform, community and user recognition and support.

charlie {l Wrote}:Your worldview is impractical. The vast majority of games made for money (as in, games that pay bills) are not open source. By using a viral license, SDL would marginalize itself massively.

The vast majority of games are made with unethical intentions, this does not make it OK to aid such activities at all -- actually the opposite conclusion should be drawn.


P.S.
I don't care one ounce about "open source".

Open source is welcome in this community.

I personally prefer open source to proprietary non-free and hope that SDL's licensing move will bring open source (rather than proprietary) software use and practices to projects that were not open to SDL under LGPL.

I'd like to note that zlib license is a free software license by all means, including FSF's point of view:

The zlib License has been approved by the Free Software Foundation (FSF) as a free software licence,[1] and by the Open Source Initiative (OSI) as an open source license.[2] It is compatible with the GNU General Public License.[1]

Sources: http://en.wikipedia.org/wiki/Zlib_License, https://www.gnu.org/licenses/license-list.html#ZLib

The GPL/LGPL licenses are not the only free software licenses for the FSF itself, but the prefered ones. A software can be free software to the FSF and not use a (L)GPL license. Indeed, the linux graphics stack has been tradicionally licensed under MIT license, and nobody complains about it, or denies to call it free software, even the FSF.

Seriously, guys, this is a non-sense discussion.

jcantero {l Wrote}:I'd like to note that zlib license is a free software license by all means, including FSF's point of view:

The zlib License has been approved by the Free Software Foundation (FSF) as a free software licence,[1] and by the Open Source Initiative (OSI) as an open source license.[2] It is compatible with the GNU General Public License.[1]

Sources: http://en.wikipedia.org/wiki/Zlib_License, https://www.gnu.org/licenses/license-list.html#ZLib

The GPL/LGPL licenses are not the only free software licenses for the FSF itself, but the prefered ones. A software can be free software to the FSF and not use a (L)GPL license. Indeed, the linux graphics stack has been tradicionally licensed under MIT license, and nobody complains about it, or denies to call it free software, even the FSF.

Seriously, guys, this is a non-sense discussion.

You are missing the point. Yes, SDL will continue to be free software. However, it will aid the development of nonfree software with its new, lax licence. «This allows people to share the program and their improvements, if they are so minded. But it also allows uncooperative people to convert the program into proprietary software. They can make changes, many or few, and distribute the result as a proprietary product. People who receive the program in that modified form do not have the freedom that the original author gave them; the middleman has stripped it away.»[0]


[0] <https://www.gnu.org/copyleft/copyleft.html>

alexander {l Wrote}:You are missing the point. Yes, SDL will continue to be free software. However, it will aid the development of nonfree software with its new, lax licence.

And the LGPL aids the development of nonfree-non-statically-linked software, too. And gcc, gmake, and the rest of toolchain can be used to compile non-free software. And the same with the rest of free interpreters and compilers. The Freedom 0 allows us to "run the program, for any purpose" and this purpose can be to make nonfree software, too.

So, we can make non-free software with free software, true, but the opposite could be not true. And because of that (and other things), free software is better than non-free software

If you ask me, I rather prefer the LGPL with a you-can-statically-compile-this-library special clausule, but a) the LGPL is very complex as it is now, it can be a real legal challenge to write such exception, b) the FSF doesn't like it.

On the other hand, being SDL simply a wrapper library, its functionality is not coded in the library itself. but the plataform below. And that includes non-free plataforms like DirectX. Hardly can anyone take advantage of free software it that case, when the real work can be done by non-free software library. And it can be seen a bit unconsequent to reclaim a free software virality clausule when we are linking agaist non-free software that is not imposing viral conditions itself.

As I said, this is a discussion with no goal, because we are in the grey areas where free and non-free software touches (a lot), and that areas only work because of compromises.

Oldest discussion ever, and it has been settled many times before.

Basically it boils down to preference, with experience shown that copy-left projects (if taken up and accepted) do show longterm overall benefits (see Linux kernel, vs FreeBSD), while "open-source" projects are quicker and easier taken up by commercial "players" (which is a good thing, but whom are sadly still scared away a bit from copy-left even though it would be basically in their interest too as it prevents competitors to take the advancements and spin-off their own even better closed source alternative).

At this point in time, open-source is all good as it enables commercial "players" to dip their feet into the free-software development style easily and any back-contribution is better than none. In the mid-term there will be some that wisen up and see the benefits of a collaborative code base, while also realize the dangers of not having copy-left code (ok, LGPL style, pure GPL is not really in the interest of commercial developers).

There you have it, argument closed!

@jcantero The toolchain was made LGPL because it was deemed necessary for adaptation at the time. The adaptation argument might have held true back then (though I think that we can today realise it was a mistake), but it is hardly relevant for SDL, which certainly doesn't have the lone goal of spreading free software.

@julius It's not a matter of preference for me, but what is ethically The Right Thing. I don't care about what's practical. Robbing someone might be a practical solution to earning money for some -- but whether it is practical or not is not the question. Adaptation, "contributing back", all of this is not relevant if it unnecessarily encourages unethical software development.

Your point about GPL not being "in the interest of commercial developers" is seemingly founded in absolutely nothing. I work with commercial GPL software. I don't see the problem.

I don't appreciate your piss poor suppression technique, but since you are a moderator I assume this is where I shut up, so this is my last post to this thread.

Na, I was rather making fun of these arguments (I could have closed the thread easily for real ).

The GPL is not in the interest of commercial developers that live purely off selling software, but you are right, there are a lot of commercial developers that have the software as a side business or as a tool they use, and for most of those the GPL is perfectly fine.

About the ethical argument (besides there being rather more pressing ethical problems in the world):
No one is being robbed, as the "open-source" developers are perfectly aware that people can take the code and use it for closed source projects. It's like putting stuff up in your lawn with a big sign, "feel free to take it, would be nice if you leave something in exchange", which is not robbery and might result in some pretty awesome stuff turning up on your lawn

Edit: about encouraging closed source development: I believe at this point in time the opposite is true. Close-source development is the defacto bad standard right now, and the GPL "encurages" people to keep developing closed-source and staying out of the GPL "walled garden" out of the fear that once they are in it they can't exit anymore (in realty that is of course not a true issue, but many current close-source developers fear that). Open-source on the other hand encourages switching to real free software coding as it as a low entry barrier for everyone, and some people might change their mind once they experienced the benefits.

That said... I am totally on your side, Copy-left is the right way of doing software development... I just don't think it is worth making an ethical argument out of it (especially when it comes to game development, voting machines are a different thing for example).

Julius {l Wrote}:Na, I was rather making fun of these arguments (I could have closed the thread easily for real ).

You can close or you can move it to a more appropiate subforum (Legal & Business?) if you want to continue with the discussion about FS vs OSS. I'm not going to insist anymore that zlib is "free software as FSF-free software definition" etc

But I wonder what are you going to do about that (if any). Forking SDL 1.2? Use another library? Use a native library? (X? but... it's MIT license) Write your own library?

Moving the licensing debate part of this thread into the Legal & Business subforum sounds like a really good idea to me. It has pretty much completely derailed the thread from its original subject.

Moderation: DONE!

Deleted.

Ahh... now that is a topic to get me started argueing In principle I agree, but a reformed patent law that is extended to what is currently covered by copyright would be actually a really good idea (and not all that unrealistic to achieve).

Deleted.

Actually, no. And if you look at the history of patents this was never how they were intended or seen (contrary to copyright, but even that was originally seen much less of a deal than today).
Patents are (or were... it's all twisted nowadays) a pragmatic solution to allow people to get the money back which they invested in research and design of a novel technical solution. It's a way of encouraging innovation.

The main positive points about (the original idea of) patents are as follows:

1. Basic research (e.g. ideas) and minor improvements of a design can not be patented.
2. By default anything is in the public domain unless you take the effort to register it with the patent office.
3. The patent office requires you to provide a nicely documented source version and stores that for reference and easy access once the patent runs out.
4. Patents are generally considered to be really only for the research investment recovery time, e.g. about 10-20 years at most.

So as you can see, the same thing could be quite easily applied to a reform of the copyright, and in a sense it is a pragmatic solution like the GPL.

But I agree, as so many things that have been created a while back as a sort of good idea (Copyrights included), the current legal system of patents leave much to be desired and need a reform urgently too.

Deleted.

FaTony {l Wrote}:I'm sorry but GPL is practical. The perfect world solution would be complete removal of copyright and patents and reform of trademark law.

What do you mean by "reform of trademark law"?

Deleted.

Julius {l Wrote}:Basically it boils down to preference, with experience shown that copy-left projects (if taken up and accepted) do show longterm overall benefits (see Linux kernel, vs FreeBSD), while "open-source" projects are quicker and easier taken up by commercial "players" (which is a good thing, but whom are sadly still scared away a bit from copy-left even though it would be basically in their interest too as it prevents competitors to take the advancements and spin-off their own even better closed source alternative).

....

There you have it, argument closed!

Oooo Julius, you couldn't let it lie, could you! That's a terrible argument. Firstly, you can not directly compare a library to an OS! They are different ecosystems.

Secondly, you make a massive and (I think) false assumption in your point; you assume it is the licensing that determined the fate of Linux vs FreeBSD. Whilst it is a factor, I'd argue it was by far the smaller factor. The licensing is just part of a development model. You have to factor in politics, requirements, developer activity (Linus was incredibly receptive for almost all of the formative years of Linux, far more so than his BSD equivalents), ease of contributing, ease of understanding the code etc etc. You've boiled a really complicated process down into a "my license is better than yours"-esque quip.

I wasn't the one that necro'ed this thread

But I stand by my point that copyleft has long term benefits and open-source/MIT like has short term benefits (and maybe long-term disadvantages).

Julius {l Wrote}:But I stand by my point that copyleft has long term benefits and open-source/MIT like has short term benefits (and maybe long-term disadvantages).

I hope you have more than just the Linux vs BSD example from which to base this on.

I would say it is more situational and less generic than you make it out to be. It has a much bigger effect for forks but even then it's not always that clear cut.

For example, LibreOffice vs Apache OpenOffice, I think the license difference is a big help here to LibreOffice and a big hinderance for Apache OpenOffice - but even then project momentum is influenced by the history and corporate politics involved. There was an attribution clause for contributing to OpenOffice which really put off contributors who were not happy that their work was explicitly exploitable by Sun/Oracle.

A better example could be XFree86 and the X.org fork - but again its just not as simple as MIT vs GPL, the change of license to the XFree86 License 1.1 really rubbed people up the wrong way.

Of course, in the event of a fork, the GPL fork is at a massive advantage over the permissively licensed. The GPL fork can absorb BSD-like licensed code but not the other way around. This gave e.g. LibreOffice and X.org massive momentum over their counterparts. Also a big factor in both examples was the unwillingness of original project maintainers to let go of their projects and allow their growing communities to take the mantle.

However you are not arguing about that kind of situation - just about how the license benefits or hinders an individual project.