GCC Will No Longer Require Copyrights Be Assigned to the FSF (devclass.com) 70
Version 9.4 of the GNU Compiler Collection "encompasses more than 190 bug fixes for GCC 9.3, which has been available since March 2020," reports DevClass.
But they add that in addition, "Developers who want to contribute to the GNU Compiler Collection but don't feel like signing over copyright to the Free Software Foundation can get busy committing now." GCC Steering Committee member David Edelsohn informed contributors via the mailing list that the committee "decided to relax the requirement to assign copyright for all changes" to the FSF. Speaking for the committee, he wrote that the GCC project "will now accept contributions with or without an FSF copyright assignment", a practice thought of as consistent with that "of many other major Free Software projects, such as the Linux kernel". GCC "will continue to be developed, distributed and licensed" under the GPLv3, so nothing should change for those adding to the project under the old assumptions.
There are those who have had troubles with that arrangement before, with Apple often cited as a popular example. They are now free to contribute utilising the Developer Certificate of Origin instead of agreeing to an FSF Copyright Assignment.
A reason was not given, though the last sentence of the statement, which affirms the principles of Free Software, might give a clue. In March 2021, the committee commented on the removal of Richard Stallman from the project's steering committee website with a similar declaration... [T]hey felt like an association with Stallman was not serving the best interests of the GCC developers and user community, given that the "GCC Steering Committee is committed to providing a friendly, safe and welcoming environment for all."
The Register notes that Red Hat senior principal engineer Mark Wielaard asked why there was no public discussion before making the change.
But they add that in addition, "Developers who want to contribute to the GNU Compiler Collection but don't feel like signing over copyright to the Free Software Foundation can get busy committing now." GCC Steering Committee member David Edelsohn informed contributors via the mailing list that the committee "decided to relax the requirement to assign copyright for all changes" to the FSF. Speaking for the committee, he wrote that the GCC project "will now accept contributions with or without an FSF copyright assignment", a practice thought of as consistent with that "of many other major Free Software projects, such as the Linux kernel". GCC "will continue to be developed, distributed and licensed" under the GPLv3, so nothing should change for those adding to the project under the old assumptions.
There are those who have had troubles with that arrangement before, with Apple often cited as a popular example. They are now free to contribute utilising the Developer Certificate of Origin instead of agreeing to an FSF Copyright Assignment.
A reason was not given, though the last sentence of the statement, which affirms the principles of Free Software, might give a clue. In March 2021, the committee commented on the removal of Richard Stallman from the project's steering committee website with a similar declaration... [T]hey felt like an association with Stallman was not serving the best interests of the GCC developers and user community, given that the "GCC Steering Committee is committed to providing a friendly, safe and welcoming environment for all."
The Register notes that Red Hat senior principal engineer Mark Wielaard asked why there was no public discussion before making the change.
Swastiki ASCII Art. Paid By The Bit. (Score:2)
Assigning copyright was weird (Score:2, Insightful)
Requiring assigning copyright to the FSF was weird. Big projects like the Linux kernel are pretty much guaranteed to stick with a GPL-compatible license because getting approval from all the contributors for a license change would be a nightmare. Assigning all the copyright to one organization means that organization can do a license change whenever they feel like it. Closed-source premium versions for example, or making the whole project closed at some point in the future.
Re: (Score:1)
Re:Assigning copyright was weird (Score:5, Informative)
it's not weird at all. when a court case occurs, having a single assignee (in particular, a Foundation whose Charter is *absolutely* guaranteed to pursue cases in the interests of Software Freedom), is extremely important. there are multiple other reasons - i don't know the full technical / legal details but have had them explained to me (a long time ago) by prominent people who know what they're talking about.
ah, i remembered one of them. let's imagine that someone is stupid enough to violate the GPL License. under the GPL they actually lose all rights to use the software. this is an instantaneous and irrevocable loss. the *only way* to get rights to use the software once you have permanently and irrevocably lost those rights? request them to be reinstated from the Copyright Holders.
now, in the case of the Linux Kernel, that's flat-out impossible. tens of thousands of contributors - or their estates - **ALL** need to give their consent.
however for copyrighted material that's been assigned to a single entity? dead easy. demonstrate that you've fixed the copyright infringement that you engaged in, ask for their review, and they will likely restore your rights to use the software.
this decision is basically by far the stupidest self-harming decision made by a Software Project in a long time.
Re: (Score:3, Interesting)
This is a term of the GPL, which the FSF can of course make whatever it wants. It does seem kind of short-sighted. It would be bett
Re:Assigning copyright was weird (Score:5, Informative)
it's not weird at all. when a court case occurs, having a single assignee (in particular, a Foundation whose Charter is *absolutely* guaranteed to pursue cases in the interests of Software Freedom), is extremely important. there are multiple other reasons - i don't know the full technical / legal details but have had them explained to me (a long time ago) by prominent people who know what they're talking about.
ah, i remembered one of them. let's imagine that someone is stupid enough to violate the GPL License. under the GPL they actually lose all rights to use the software. this is an instantaneous and irrevocable loss. the *only way* to get rights to use the software once you have permanently and irrevocably lost those rights? request them to be reinstated from the Copyright Holders.
I don't think that's correct. FSF wants the copyright because having a clear copyright on the entire code base makes it very straightforward to sue a violator [gnu.org].
As to losing rights the GPL automatically reinstates once violations cease unless a Copyright holder pro-actively terminates the license: [gnu.org]
However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation.
Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice.
Now I don't know what happens if some of the copyright holders try to permanently terminate the copyright. That's a very hypothetical scenario but I suspect that you need ALL of the copyright holders to agree to give you a license to use the code, so giving the FSF an additional copyright on a file doesn't change anything here.
now, in the case of the Linux Kernel, that's flat-out impossible. tens of thousands of contributors - or their estates - **ALL** need to give their consent.
Wrong, the Kernel has a lot of big copyright holders with big legal teams (IBM) who could easily sue for violation. The Kernel didn't do the FSF copyright assignment thing, OR include the GPL v2 clause that automatically updated to new versions of the GPL, because they didn't trust the FSF (nor I suspect, would various coprorate legal departments). Given the success of the Kernel project including contributions from big companies I think this was a wise choice.
this decision is basically by far the stupidest self-harming decision made by a Software Project in a long time.
Actually, it's probably one of the least relevant legal decisions made by a Software Project in a long time. The FSF will still have an assignment on the vast majority of the code base, and I don't see any practical consequence on them lacking copyright on a handful of files.
Re: (Score:2)
I don't think that's correct. FSF wants the copyright because having a clear copyright on the entire code base makes it very straightforward to sue a violator [gnu.org].
As to losing rights the GPL automatically reinstates once violations cease unless a Copyright holder pro-actively terminates the license: [gnu.org]
this is the case for the GPLv3. it's _not_ the case for the GPLv2. under which the linux kernel is distributed. the linux kernel is GPLv2 *only*, not GPLv2+. this is not a decision that can in any practical way be rescinded, because tens of thousands of copyright holders (or their estates) now need to be contacted, to, as a group, change that decision and issue a new license. ... start to see why having a single copyright assignment is important?
Re: (Score:2)
Yes... it's super weird. GCC devs aren't paid for their work. I hope they will also provide a way for developers to Cancel their copyright assignments and get their ownership back as well. They should do it to make sure developers don't have to wait 35 years before they can invoke their right to revoke [authorsguild.org] all copyright assignments/transfers.
Re: (Score:2)
> Yes... it's super weird. GCC devs aren't paid for their work.
Most are paid by their workplace, even if it's student work. The workplace retains copyrights to their work: putting a copyright in the hands of the FSF was useful protection against confusion or abuse by an author or organization changing policy and withdrawing their copyrighted work, especially as part of a much larger work like gcc. Since Richard M. Stallman _wrote_ gcc 1.0, it seemed especially reasonable to put the copyrights in the hand
Re: (Score:2)
Most are paid by their workplace, even if it's student work. The workplace retains copyrights to their work
Last I check "student" is not an employed position, unless that person happens to be a Teaching Assistant. Even then, the Student would still own the copyright.. Workplace is not a thing that gets rights; Employers can claim copyrights of their employees, but only if it's work made for hire by Employee or a work specially commissioned under a contract to make the work for the employer - the work
Re: (Score:2)
You've a point that employers retain copyrights, not workplaces. However, few undergrads today are capable of contributing to gcc. _Grad_ students are more likely to do so, and _they_ are likely to be working on something for their adviser that belongs to the university. Whether work on gcc that is published in binary form and sent to others is an interesting issue. Many corporate lawyers consider the GPL very dangerous to their business models, and it can be, when what they wish to sell is proprietary acce
Re: (Score:2)
Grad student IP typically does not belong to the university. Copyrights generally belong to whoever wrote it.
Graduate students are typically not "employed" by a university. They pay tuition and may receive scholarships, stipends, bursaries, TAships, or some combination. TAships are specific engagements to teach a class, not to write code.
Re: (Score:2)
But ...
You are fully aware that "copyright" does not work like that outside of the united states?
For starters: in most countries you can not "assign the copyright" to someone else.
And the reason why some Gnu run projects want me to "assign my damn rights" to them, is why I never contributed. By putting it under GPL, I granted them already what they need: the right to do what they want - in reasons. Why they want me to GIVE UP MY OWN COPYRIGHT by assigning it to someone else: was always beyond me.
Re: (Score:2)
Which "most countries" are these? Because they're not countries in the European Union [europa.eu]:
Re: (Score:2)
That's the weirdness right there. The point of the GPL is that it grants sufficient rights for you to do anything with my code that is compatible with the GPL. Assigning *all* rights is not only unnecessary, it's against the spirit of the GPL.
Re: (Score:2)
Exactly.
And the argument, that an organization like the FSF can only sue if you assign copyright is wrong in most jurisdictions.
Under the Bern Convention, every single contributor can sue for any infringement regardless if it is associated with a file *he* himself has contributed or anyone else.
Under copyright law it is a "joined work", and everyone participating can sue on behalf of everyone else.
Re: (Score:2)
It seems like enforcement would be much more effective if there were a whole bunch of copyright holders, any of which could sue, instead of just one who accepts large donations....
Re: (Score:2)
As I said: in Europe, particularity Germany anyone holding a piece of "copyright" can sue for infringement on behalf of everyone. After all the "whole work" got infringed and not a particular part of it.
Re: (Score:3, Interesting)
No, I think it was fully consistent with the goals of RMS and many others in the FS movement. The goal is to pull software in to the GPL licensing regime, and keep it there,. As long as developers retained their own copyrights, they were free to incorporate their works in to other projects under other licenses, even proprietary ones if that suited their needs. This wouldn't apply for simple patches, but would certainly have implications where the "patch" was separate files that got linked in.
Without copy
Re: (Score:2)
Lol. I can't tell who's modding this overrated: GPL zealots or GPL haters. Maybe the Free Stallman lobby?
GCC no longer free (Score:5, Interesting)
Re: (Score:1)
No mod points, so people mod him up. That is my concern, another ploy for Corporations to take control of FSF and change the license to something that allows full lock-down on us.
Maybe because the failed to remove RMS
Re: (Score:2)
i fail to see how this move can ever be used to change the license. the dco expressly states the modified code uses the same license as the project.
what this probably can do is partially restrict the options of the fsf in (a remotely possible) case of litigation, as they'll be no more the sole copyright holder. that indeed isn't good, but i don't think it is the end of the world either.
not at all an expert in license legalese and i don't really understand this move, but i don't see reasons for doom calling
Re: (Score:3)
This move would make it harder for anyone to change the license terms. That's one of the reasons the Linux kernel expressly didn't require copyright assignment. With multiple copyright holders you need the agreement of all copyright holders to change the license terms. If any copyright holder declines to change, then the license on the whole can't change without first going in and removing all of that copyright holder's code. And if any copyright holder changes the terms on their code to be incompatible wit
Re: (Score:2)
With multiple copyright holders you need the agreement of all copyright holders to change the license terms. If any copyright holder declines to change, then the license on the whole can't change without first going in and removing all of that copyright holder's code.
Agreement would be explicitly granted by the fact the code was submitted to the GCC which is under GPL.
And if any copyright holder changes the terms on their code to be incompatible with the existing license on the whole, it has no effect on code they contributed before the change attempt because of the way the GPL is worded (eg. GPL v2, paragraph 4, and more explicitly in GPL v3 paragraph 2).
Courts generally do not like it when contract agreements are changed without the consent of both parties in the agreement. By submitting code under the GPL (or any license), an agreement was reached. The party submitting the code simply cannot change their mind in the future without the consent of the other party.
Re: (Score:2)
Using the GPL for your code definitely does not grant a right for others to change the license on your code.
As for the courts, remember that a license isn't a contract. There is no negotiation, there is no consideration, it's a unilateral grant by the copyright holder of permissions, subject to certain conditions, that the recipient would not otherwise have. That's already been litigated and ruled on at the federal appeals court level (see Jacobsen v. Katzer [wikipedia.org]) so it's binding precedent in all copyright cases
Re: (Score:2)
Using the GPL for your code definitely does not grant a right for others to change the license on your code.
My point again: A person submitting code under the GPL cannot retroactively decide they can change the terms of that license. Someone could release future versions of code under a different license but not the existing code.
As for the courts, remember that a license isn't a contract. There is no negotiation, there is no consideration, it's a unilateral grant by the copyright holder of permissions, subject to certain conditions, that the recipient would not otherwise have.
Software licenses are implied contracts [investopedia.com] but they are contracts.
That's already been litigated and ruled on at the federal appeals court level (see Jacobsen v. Katzer [wikipedia.org]) so it's binding precedent in all copyright cases.
How does Jacobsen determine that software licenses are not contracts? What binding precedent was set in Jacobsen? Your own link mentions nothing of the sort: "The case addressed claims on copyright, patent invalidity, cybersq
Re: (Score:2)
Re: (Score:3)
No it isn't. It's still being released under the GPL when they submit. What it may do is make it slightly harder to sue GPL violators of GCC, as you'd need to have the person who owns that copyright file the suit. But it doesn't make GCC any less free. Please note the majority of GPL projects to not require copyright assignment.
Re: (Score:1)
Re: (Score:2)
Copyright retention is the norm for FOSS projects. Can you cite any examples where rights were clawed back after a submitter knowingly contributed to FOSS?
Non-exclusive licensing happens all the time. We're soaking in it. I've granted Slashdot the non-exclusive license to use comments for over 20 years now. I can't sue Slashdot for displaying my copyrighted works.
Re:GCC no longer free (Score:5, Informative)
The Berne Convention considers the Author's Right a natural right, which can not be attributed to anyone else than the original author. A "work for hire" does exist, but only in the sense that the author has a contract with the hiring entity in which he assigns certain exclusive derivative rights (like the right to publish) beforehand.
Re: GCC no longer free (Score:2)
Re: (Score:3)
Re: (Score:2)
It doesn't give me any other rights.
Yes it does.
It gives you, and only you, the right to sell, publish copy, distribute and exploit your own work in anyway you want. No one else can use your work for anything, unless you give him explicit permission.
Re: (Score:2)
Not quite. Only moral rights cannot be assigned [cornell.edu]. Economic rights, including the right of reproduction, are expressly transferrable.
"(1) Independently of the author's economic rights, and even after the transfer of the said rights, the author shall have the right to claim authorship of the work and to object to any distortion, mutilation or
Re: (Score:2)
Economic rights, including the right of reproduction, are expressly transferrable.
Nope. They are _assignable_, but not transferable. They remain with the author.
and only if you enumerate them. Any rights not covered by the assignment remain with the author, because cover-all terms like "all derivative rights" are not legally valid.
Dead wrong. That's not specified by the Berne Convention, and not a requirement of, for example, U.S. law. Wherever you're claiming such a feature exists, it's a creation of natio
Re: (Score:2)
I already provided the link and the language, where the treaty expressly says transfer [cornell.edu]: "Independently of the author's economic rights, and even after the transfer of the said rights, the author shall have the right to claim authorship of the work and to object to any distortion, mutilation or other modification of, or other derogatory action in rela
Re: (Score:2)
I think you are nitpicking about a thing no one talked about.
Copyright != moral rights/author rights. You can not transfer authors rights/moral rights.
Yes, you can "transfer" economical rights based on authors rights.
But not in the way you try to make us believe.
So: no idea why you want to nitpick about a legal topic where you are not a legal expert in.
In American copyright law, if one simply "transfers copyright", all the authors rights are void. Obviously you can make a perfect contract describing perfect
Re:GCC no longer free11 (Score:2)
I think that you talked yourself into an error and are trying to recharacterize what was being talked about in a failing attempt to save face. Let's look back to the beginning:
Re: (Score:2)
Again, you are nitpicking about stuff no one argued about.
Copyrights do not exist in Europe. Hence they are neither transferable nor assignable.
It is called authors rights or oral rights. Which are not transferable.
What is possible is to "assign" (or in limits even transfer) economic rights.
And now put your head back into your ass for nitpicking stupid bullshit.
Re: (Score:2)
So not only are you an idiot, you live in a different reality [europa.eu] from the rest of us.
Copyrights exist in Europe [wikipedia.org]. We're done here.
Re: (Score:2)
You got really quiet about this one. Did you finally discover German copyright law [gesetze-im-internet.de] in the last 24 hours?
Re: (Score:2)
In olden times, I wrote Unix code at a NASA research center.
It was picked up by Berkeley, FSF, and assorted different companies.
(I date myself with the phrase "real men program in C".)
As a (somtimes-co) author and free-spirit, I always said, sure, if
you insist, I'll sign anything (and did), but what did it really mean?
Like, do programmers even have "legal" authority do even do this?
My NASA boss didn't even care, because government work is
basically public domain. If code had certain value, it was supposed
to
Re: (Score:2)
Because if copyright is not assigned to FSF, then it remains with Microsoft, Apple, Oracle, Amazon, whoever submitted. Basically a suicidal move.
Meaning what?
I don't believe you understand how the GPL works [gnu.org].
Re: GCC no longer free (Score:2)
What the FSF canâ(TM)t do is license the whole of gcc say to Google with GPL-incompatible terms.
Re: (Score:2)
Re: (Score:2)
Because if copyright is not assigned to FSF, then it remains with Microsoft, Apple, Oracle, Amazon, whoever submitted. Basically a suicidal move.
If this were a problem it would completely invalidate the entire point of the GPL, which is of course is the FSF's single greatest achievement.
Re: (Score:2)
Because if copyright is not assigned to FSF, then it remains with Microsoft, Apple, Oracle, Amazon, whoever submitted. Basically a suicidal move.
So basically the Linux Kernel has commited suicide? I feel like you don't really understand what "copyright" is or how open source licenses work. GCC was a complete odd ball with this requirement, and not having the requirement doesn't make it's open source license any less applicable.
Re: (Score:1)
It only makes changing license on gcc harder.
Disgusting (Score:1)
We need to get rid of these disgusting bullies before they destroy everything they touch.
Conspiracy (Score:2)
Re: (Score:2)
Re:Conspiracy (Score:4, Insightful)
No, the connection is that for developers to feel comfortable assigning copyright to the FSF they need to trust the FSF, and that trust has been eroded recently. People on both sides - those who criticize Stallman and those who support him - are upset with the way the FSF handled the situation.
There are likely other reasons for the change beyond this as well. There are practical concerns around the fact that the FSF has been taking a very long time in processing copyright assignments. There are cases where it has been taken over a year [gnu.org], during which time commits are held up from being merged waiting for FSF bureaucracy. There are also the (far less common, almost hypothetical) issue where someone would like to contribute code to multiple projects, and if they both require copyright assignment, that is simply not possible. Lastly, some people have always object to the principle of assigning their copyright, whether they trust the FSF now or not.
That said all of these are general reasons why GCC might have chosen to change their policy, they haven't made their own reasoning public.
Re: (Score:3, Informative)
No, the connection is that for developers to feel comfortable assigning copyright to the FSF they need to trust the FSF, and that trust has been eroded recently.
this is false on two counts:
1) it's a false statement in itself. a correct statement would be, "efforts by the press and those interested in taking control of Software Libre have *portrayed* the FSF as untrustworthy"
2) you also demonstrate a very common misunderstanding about the *concept* of trust. Trust is *GIVEN*. it's never *EARNED*.
the conflation is best illustrated with the phrase "you have to EARN my trust". using this attitude and weaponising it is how you take down an enemy: "earn" their trust.
Re: (Score:2)
No, the connection is that for developers to feel comfortable assigning copyright to the FSF they need to trust the FSF, and that trust has been eroded recently.
this is false on two counts:
1) it's a false statement in itself. a correct statement would be, "efforts by the press and those interested in taking control of Software Libre have *portrayed* the FSF as untrustworthy"
No, that would be a statement of opinion. A correct statement of fact is exactly what the previous poster made.
precisely. total lack of transparency. do you TRUST an organisation that is not transparent?
You mean like the FSF re-instating Stallman to the board with zero notice (even to their employees) or consultation?
That is in fact one of the reasons I don't trust the current leadership of the FSF.
Re: Conspiracy (Score:2)
Re: (Score:2)
Is this article really suggesting that Stallman was forced out of the FSF due to open source licensing disputes with Apple?
No, various people and groups long thought the FSF copyright assignment was weird, Apple being an influential example, but no one did anything about it since the FSF was viewed as either positive at best or irrelevant at worst.
But now after all the trouble with Stallman many now view them as objectionable and unreliable, and requiring contributors to assign copyright to an organization they find objectionable and unreliable is a problem.
If developers keep the copyrights (Score:2)
Re: (Score:3)
The GPL.
Re: (Score:2)
Re: If developers keep the copyrights (Score:2)
Re: (Score:2)
The point of irrevocably granting a license is that you can't revoke it.
Re: (Score:2)
Nothing is free (Score:2)
Wanting my copyright assigned is _not_ a fair deal. Never was. If I figure out a new, improved register allocation algorithm, contributing it to gcc under the GPL license is fair. Assigning the copyright, which