Is Apple Now Blocking Contributions To GCC?

[nanonyme]

Not that this affects most of you lot who are so busily arguing about it.

[RealNC]

The greed here is with FSF; they're greedy and want more than they deserve.

The code is there. Either take it or leave it (just like Linux does.) Apple is in no way required to play ball with FSF's overly anal policies.

The FSF fights for more open source code, and now that they get it, they want more than just open source: they want to *own* all the code. If that's not greed, then I don't know what is.

[allquixotic]

The greed here is with FSF; they're greedy and want more than they deserve.

No; they just want to remain consistent with policies that have existed since before Objective-C as a language existed. They have always (rightfully) been concerned about the consequences of having their code copyright attributed to other entities, because the ability to relicense the code quickly approaches impossibility as the number of copyright assignees increases.

This isn't about GPLv3 or any particular license. It's about the principle of being able to relicense code as the legal landscape changes. Not even proprietary companies are unrealistic enough to believe that a single, static work of text can remain legally solvent and relevant over an indefinite period of time in the face of ever-changing laws in jurisdictions around the world. That is why many EULAs change from year to year in subtle ways. Since the GPL isn't a form of WTFPL (where the answer to a question beginning with "Can..." is "Yes!"), the license text necessarily must evolve over time, even if all they want to do is maintain exactly the same level of freedom that is granted today. But Linux, for instance, is something that will never see the benefit of a new license, even if it were something like, hypothetically, a GPL v2.1 that just clarifies some legal language to improve its defensibility in court, without adding any new clauses.

Being stuck with the same license in perpetuity is a really foolhardy thing to do, regardless of whether the copyright holder's intentions are pure as snow, sinister as MSFT, or something in between. The disadvantages are plain to see, and they only become more severe as time goes on. And the only way to guarantee that you will be able to relicense the code is to keep the number of copyright assignees down to a manageable number -- less than 10 would be practical; 2 to 5 would be easy; but 1 is the easiest and most guaranteed route.

And for the record, there are many other projects and companies that require copyright assignment in order to contribute -- Ubuntu, for example. I don't think it is necessary to have all FOSS code licensed to the FSF; rather, that each separate software project/product has its own copyright assignees kept to a manageably low number. Once the number exceeds 1, the chances of disagreements, death, etc. increase, so really the best thing is to keep it to 1.

The code is there. Either take it or leave it (just like Linux does.) Apple is in no way required to play ball with FSF's overly anal policies.

It's only "overly anal" if you can't take the time to understand why they do it that way, just like any rule or restriction. Is the US Government overly anal about safeguarding nuclear resources that could be weaponized? Is GregKH overly anal about not allowing potentially breaking changes into the stable kernel updates? The world doesn't operate smoothly on the basis of anarchy. Even in the free world, there has to be an order to things.

The FSF fights for more open source code, and now that they get it, they want more than just open source: they want to *own* all the code. If that's not greed, then I don't know what is.

No, the FSF fights for more Free Software code, not open source. Just correcting your terminology. Don't put words in the FSF's mouth. You can call Free Software and Open Source the same thing when you're talking generically about code falling under a license that is both Free and Open Source, but when you're saying what the FSF fights for, at least represent what they fight for, accurately.

The GNU project's copyright attribution rules are nothing new. They have always existed. Everyone who has ever contributed to a GNU system has known about these rules. Has the FSF always been greedy? What makes them greedy now, if they weren't greedy in 1985 when they were trying to start a movement to promote the sharing of knowledge? Open Source didn't exist back then, but the copyright attribution policy did. I ask you, at what point does an unchanging policy become greedy when it wasn't before?

Seriously, go use LLVM and BSD and BusyBox if you don't like (or more commonly, don't understand) GNU's policies. The onus is, and always has been on contributors to be aware of the attribution policy and to accept it. This has worked out for the FSF and GNU since I was born. One proprietary software company decides to go against the policy and suddenly GNU is the bad guy? Maybe that's exactly what Steve Jobs wanted; to turn people like you against the Free Software movement. You're playing right into their hands.

[deanjo]

No; they just want to remain consistent with policies that have existed since before Objective-C as a language existed.

Ummm, Objective-C actually predates GCC.

[deanjo]

Also open source had been long around before the FSF and the Open Source Initiative.

[RealNC]

No; they just want to remain consistent with policies that have existed since before Objective-C as a language existed. They have always (rightfully) been concerned about the consequences of having their code copyright attributed to other entities, because the ability to relicense the code quickly approaches impossibility as the number of copyright assignees increases.

Relicense? Is GPL version X or any later version not good enough? You know, by keeping the code's copyright to myself, I am effectively disallowing someone else to relicense it without my permission.

If the FSF won't accept that, well, it's their loss. The GPL was invented for a reason. So if the GPL is now not good enough for the FSF, then I don't know what to say.

Seriously, go use LLVM and BSD and BusyBox if you don't like (or more commonly, don't understand) GNU's policies. The onus is, and always has been on contributors to be aware of the attribution policy and to accept it.

Apple is not a contributor in this case. They simply comply with the GPL, as the FSF has envisioned it.

This has worked out for the FSF and GNU since I was born. One proprietary software company decides to go against the policy and suddenly GNU is the bad guy? Maybe that's exactly what Steve Jobs wanted; to turn people like you against the Free Software movement. You're playing right into their hands.

It's hard not to; the code is GPL, everyone is "free to share, modify and redistribute it." It simply looks like the FSF is whining about a non-issue.

And besides, Linux kernel development has worked out too, and they don't even have the "any later version" clause.

[deanjo]

Relicense? Is GPL version X or any later version not good enough? You know, by keeping the code's copyright to myself, I am effectively disallowing someone else to relicense it without my permission.

If the FSF won't accept that, well, it's their loss. The GPL was invented for a reason. So if the GPL is now not good enough for the FSF, then I don't know what to say.

Bingo. Hit the nail on the head.

Apple is not a contributor in this case. They simply comply with the GPL, as the FSF has envisioned it.

Absolutely true, it is the FSF that is tightening the noose around their own neck by adding more and more restrictions to itself.

[allquixotic]

Ummm, Objective-C actually predates GCC.

But it doesn't predate the copyright attribution policy of GNU.

"In 1986, Cox published the main description of Objective-C in its original form in the book Object-Oriented Programming, An Evolutionary Approach." from Wikipedia

"It initiated the GNU operating system, software development for which began in January 1984." from Wikpedia

Regardless of the actual timing surrounding the issue, I don't think it is relevant to the discussion. I was simply using the rhetorical comparison to illustrate how the project (and the policy) has been around the block for a couple decades, so it shouldn't be a new and surprising fact to people.

Normally I, too, derive some concern from compulsory copyright attributions, but I limit my concern to for-profit companies. As we have seen with OpenSolaris, for-profit companies have a financial incentive to open up their code base, allow a whole bunch of contributions to get attributed to them, then close it back up again. OpenOffice.org is, unfortunately, now in the same scenario, and it may be next on Oracle's chopping block of betrayal. Although I judge Canonical as a much more altruistic company overall, I do indeed have my concerns (although they're pretty far wedged in the back of my mind) that their altruism may one day run out. They are in it to make money, after all.

But the FSF can't sell out, be bought or taken over like a for-profit company; as long as Stallman or followers of the Stallman school of thought run the FSF, it is unfathomable to me that they would ever do something as sinister as Oracle has done. If you're that paranoid, you would probably also believe that Stallman works for Microsoft, and that pigs can fly. If that's the case, I'll bow out; you can't argue with a madman.

Copyright attribution is a double-edged sword: sometimes it makes sense, other times it doesn't. I would be perfectly fine with all Linux kernel contributions getting attributed to the Linux Foundation -- another non-profit -- or all GNOME contributions getting attributed to the GNOME Foundation. This isn't reality mostly for historical reasons, I think. Also consider that the GNOME Project consists of many separate programs and libraries, each of which is a separate software entity with its own license of the author's choosing, so it is not quite the same deal as the Linux kernel. It would actually be easier for individual GNOME projects to attribute their copyrights to the GNOME Foundation, because the number of contributors on some of the projects is extremely small.

The main purpose of copyright attribution is to facilitate license upgrades (think of licenses like software; imperfect works that need to be continually improved, just that licenses are improved over a larger time scale). The secondary purpose is to establish solidarity among a base of contributors, by having each contributor explicitly say, "Yes, I am on board with this project's mission; yes, I fully support the goals of this project and am willing to support their decisions in the future". The laissez-faire licensing of projects without a central, non-profit, attribution entity will ultimately be more problematic in the end.

If contributors on the same project can't even agree what license the project should be under, they have no business contributing to it at all. Why would anyone contribute to a project whose license they disagree with? Right... so let's see how the two different models cope with this.

In the laissez-faire model (e.g. the Linux kernel), anyone who disagrees with the original license under which the project was founded can never become a contributor. There is no way the project could ever be dual- or tri-licensed, and there is no way to satisfy the people who decline to write software under the original license. The license is set in stone precisely because the number of contributors makes it a statistical certainty that someone will stonewall any potential license change, even if 99/100 people agree that it would be extremely beneficial.

In the central attribution model (e.g. the GNU project), the potential for dual or tri-licensing, or a change of the original license, is always within reach. That means you can pick up new contributors who particularly agree with the new license but not the old. You may also lose contributors who agree with the old license but not the new; but this is only likely to happen if you pull an Oracle. People who -- by definition -- have explicitly attributed their copyrighted work to you in the past are doing so while being explicitly aware of what they're doing. They do it because they believe in the mission of the central organization. So if the central organization turns around and says it wants to use a new license, contributors are likely to support them, or at least not to hate them for it. This reasoning has held up for every non-profit that I've heard of that acts as a copyright holder for software.

I think it is overly-cynical and shortsighted to say that we should explicitly use and exploit the fact that a license can't change when it gets held up in copyright limbo for a large project like Linux. This situation may "protect" Linux from being able to get relicensed under some license you may not like, but at the same time, it keeps Linux on legally flimsy ground, which will gradually slip away from them as the law evolves and the license doesn't. The GNU project software will not suffer the same fate, so they will win out in the long term, as long as the intentions of its management remain as they have since the year I was born.

[deanjo]

But it doesn't predate the copyright attribution policy of GNU.

--snip--

long term, as long as the intentions of its management remain as they have since the year I was born.

That's all nice and all but IMHO that is still asking to "put all the eggs into one basket", entrusting it to one person who continuously changes the rules of the game. It no longer respects the rights of the contributor without asking them to fall on the sword first.

[RealNC]

What's ticking me off here is the title: "Is Apple Now Blocking Contributions?" Why does no one talk about that? Yes, Apple is blocking contributions. Not the FSF by not wanting to take the code that is there. Apple has the power to block contributions.

Seriously, this is so obviously wrong, it should even be needed that someone explains it.