[Fedora-legal-list] Change in classification of CC0
lwn.net
lwn.net
If instead the real reason is Fedora wants all licenses to include explicit "patent licensing or patent forbearance", that's different than the rationale excerpted by LWN here.
In a situation like this where the license explicitly says "this is not a waiver of any patent rights" that argument is going to be pretty hard to make fly.
The whole area is interesting for sure. The very presence of explicit patent grants/waivers/forbearance in later licenses could be argued as evidence that a waiver isn't commonly understood to automatically come with the license where it's totally silent. But that's still a different world from the "definitely not a waiver" situation set up by CC0.
This feels incompatible with a spontaneous patent claim, because otherwise "use" is not permitted once a patent right is asserted.
The notion of open source is under threat from patents being used by the likes of Facebook, who've released BSD code then immediately claimed a special patent license in a parallel that disavowed "redistribution and use" unless the patent license also applied.
So we can say that a license that explicitly doesn't include patent permission doesn't meet the 4 freedoms.
> The freedom to run the program as you wish, for any purpose.
A patent overrides this.
> The freedom to study how the program works, and change it so it does your computing as you wish.
A patent overrides this, at least the second half.
> The freedom to redistribute copies so you can help your neighbor.
A patent overrides this.
> The freedom to distribute copies of your modified versions to others.
A patent overrides this.
Facebook's React library was an example of this. BSD but with a non-free patent license that ran against those freedoms. So it was not really open source.
If the patent holder puts a "redistribution and use are permitted", that can be deemed as an IP license in every relevant way, including patent and trademark grant; it's not just copyright.
"Use" means carrying out the patented functionality; permission to "use" is tantamount to permission to apply the patented algorithm to one's data to reap its benefit.
Perhaps the Creative Commons lawyers erred by mentioning patents at all, and they should issue a corrected “CC0v2”!
Whereas with CC0 it's explicit that the same person who granted you the code could sue you for running it. There's no room for interpretation.
But it does do exactly that.
> It just makes clear that CC0 is silent on those issues
It doesn't do that, because that's impossible to do. Saying "I'm not talking about X" is talking about X.
Someone can absolutely say “Though some might think my prior words endorse or reject X, let me be clear I don’t endorse or reject X.” You can mention X, especially in a aspires-to-be-rigorous context like law, without committing to a treatment.
Yes, it does.
A license that gives permission to use software but does not enumerate specific rights that the licensor holds which may be necessary to use the software will most naturally be read as licensing all necessary rights held by the licensor.
A license which explicitly disclaims any grant of patent rights can only be read as...not granting patent rights held by the licensor, even if they may be necessary to use the software. Thus, such a grant which would otherwise be reasonably included is, instead, explicitly precluded.
Yes, and this seems to be the entire reason that Fedora is revoking approval of the license.
"The reason for the change: Over a long period of time a consensus has been building in FOSS that licenses that EXPLICITLY DISCLAIM PATENT FORBEARANCE cannot be considered FOSS."
...I'd have not flagged any apparent textual contradiction – and understood the rationale better. It was the broad, unqualified allegation that CC0's text would "preclude any form of patent licensing or patent forbearance" that created the impression of mismatch between reasoning and literal text.
(And note: submission link has changed since I commented from a tiny excerpt of Fontana's announcement, and the full email.)
I dont think it is. I really dont see how you could read that sentence and get "i reject x" out if it.
More specificly its saying "the previous statement i just made is not me saying i am giving you a patent license. However i'm not saying in general whether or not i am going to give you one, and i might be giving you one in some other statement."
There is of course a certain sense where not rendering an opinion is the same as the supporting the status quo ante (which would be no patent license), but i dont think its fair to equate an implicit no opinion = no patent license with an explicit no patent license.
> In most of Europe, it is not possible for authors to assign or even waive their moral rights. This is following a tradition in European copyright itself, which is regarded as an item of property which cannot be sold, but only licensed. The author can agree to waive them to a limited extent (and such terms are very common in contracts in Europe).
(from https://en.wikipedia.org/wiki/Moral_rights#In_Europe)
IIRC, moral rights are also always tied to the individual and not to the employer or organization.
In NL, these moral rights include the following properties:
> even after transferring their copyright, the original creator retains the rights to:
> 1a. oppose publication of the work without attribution
> 1b. oppose publication of the work under a different name than their own (applies to both the name of the work and the name of the creator)
> 1c. oppose modifications of the work
> 1d. oppose any kind of mutilation of disfiguration of the work, or modifications that would negatively affect the perception of the work or its creator
> Of these rights, 1a can be waived entirely; 1b and 1c can be waived only with regards to the name of the work, not its creator; right 1d remains the privilege of the creator (or any designated heir) for the entire duration of the copyright itself.
(from https://wetten.overheid.nl/BWBR0001886/2015-07-01/#Hoofdstuk...)
edit: note that (1c) talks about modifications of the original (physical) work; it doesn't necessarily apply in the same sense as the GPL, where all extensions/updates to the original codebase are also considered modifications of the same work.
IANAL
The problem with Public Domain is that it is decided by law, not by you. You can't put anything on the public domain; the public takes domain by expiring copyright.
> This software is information. It is subject only to local laws of physics.
I deliberately use the term "information" here to convey a sense of factual truth which is generally not protected under copyright. It is meant to be a license that transcends all human legal systems since I cannot guarantee any protections they would provide.
I have been told this will not stand up in a court of law, but the people who told me this don't seem to understand its purpose or effects. As long you obey the laws of physics no legal intervention is ever required. And the best part is that even the people who disagree must always conform to its terms.
"this software is information" does not achieve this because information can be subject to intellectual property laws! Even excluding that, it's not clear so imposes a legal risk on users.
"It is subject only to local laws of physics" does not achieve a clear grant of permission. It's phrased as a statement of fact, and a statement of fact doesn't actually give permission for a reader to just act as if its true.
There's minor issues too... just use the 0BSD or MIT-no-attribution or CC0 license which does what you want in clear and unambiguous legal language, with the bonus of trying to stop you getting sued.
How can I create unambigous legal language if I don't know the laws to which I am subject? Am I supposed to assume US or "Western" style copyright? Will this work in China? What about Russia or North Korea? Does this language work in 2122? 2222? Under what authority?
And since I live in the US, let's be clear: nothing will prevent someone from suing you. The best chance you have is to hope the judge agrees with your interpretation of the law.
- [Unlicense](https://unlicense.org/)
- [What The Fuck You Want](http://www.wtfpl.net/)
- [Public Domain](https://creativecommons.org/share-your-work/public-domain/)
- [0BSD](https://opensource.org/licenses/0BSD)
You can find more on tl;dr:legal
This has not AFAIK been tested in any court, but (to my mind) it's not an unreasonable point to accept.
CC0 actively does not provide any protections with regards to patents, which would make the above defense inapplicable.
> Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use …
It would be interesting to argue that, yes, I gave you the right to “use” this software but you actually had to pay me extra for a patent license.
MIT, BSD, and GPL 2.0 also do not surrender those rights. They just don't helpfully remind the licensee of this fact in the body of the license itself. The net result is, so far as I can tell, the same.
It would seem to me that the complaint is with a license which doesn't surrender rights, particularly patent rights. I think this is proper: MIT, BSD, and GPL 2.0 are awful licenses for exactly this reason.
*Edit* A quick mention of MIT. There are people who argue that MIT's "without restriction" clause sort of acts as a patent release. As if that was good. MIT is a bad license because a critical feature of a modern license -- a patent release -- hinges on a strained interpretation of a vague, easily misconstrued statement, and one which was clearly not the intention of the original license authors. A good license would have explicitly spelled things out. It takes just one judge to grind the MIT license into dirt.
> Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software
It doesn't say anything about copyright or patents. It merely provides a list of things that I am allowed to do. To decide that they haven't granted all rights required to use the software would seem to contradict what was written in the license.
> No trademark or patent rights held by Affirmer are waived, abandoned, surrendered, licensed or otherwise affected by this document.
and a license with no clause at all addressing trademark or patent rights? Wouldn't both of those documents leave trademark and patent rights unaffected?
> The MIT License unquestionably has an express [patent] license. That license is not limited to the granting of any particular flavor of intellectual property rights. The statement of license does not use the word "patent" or the word "copyright." When was the last time you heard someone expressing concern that the MIT License merely had an implied copyright license?
CC0 is a public domain dedication, with fallback license for jurisdictions where public domain dedication is not possible. The public domain dedication part is not a license (this was a large part of the legal criticism of the awfully-named Unlicense: that it tried to be both a public domain dedication and a license, despite the two being incompatible), and so cannot do anything with trademark or patent rights.
If you want to cover trademark or patent rights, you’ll have to use a public-domain-equivalent license instead of a public domain dedication. This is where things like 0BSD come in. (With this latest thing, I think I’m going to change my primary recommendation in https://chrismorgan.info/blog/unlicense/ to 0BSD rather than CC0.)
"Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:"
Can you use and redistribute it or not?
"Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions"
Are you free to deal in, redistribute, sublicense and sell the software, or are you not?
Fedora is ultimately stewarded by Red Hat, which (let's not forget)... is now an IBM company. As such, a cynic might be inclined to read this notice more like "IBM doesn't like FOSS licenses that are (or might possibly be construed as) patent-prohibitive." At the risk of sounding churlish... NO $**T! Phrased more delicately, methinks one's motives might be suspect.
I find this little more than another pawn moved around the board in the whole "how cozy with big business should FOSS be" debate, and my personal take is that we passed most sane thresholds a decade or more ago. IBM via Red Hat via Fedora can 2>&1>/dev/null .
List of Germans who've been sued for using SQLite without a license: .... [End of list]
Build a billion dollar company on some "public domain" software in Germany, and you've given a stranger a gun to your head. They might not exercise their copyright, but they also could.
Legally that should work, even for non-Americans: the 0BSD doesn't grant anything the CC0 doesn't. But that doesn't work if you assume the 0BSD has a patent grant (by arguing that it gives you permission to use, modify, etc. the software). Then the product of your license reassignment is 0BSD, but doesn't grant permission from all authors, thus preventing its use in Fedora.
If I search for CC0, the first hit is https://creativecommons.org/share-your-work/public-domain/cc... which already starts to answer from the URL, but let me just quote it:
> CC0 enables scientists, educators, artists and other creators and owners of copyright- or database-protected content to waive those interests in their works and thereby place them as completely as possible in the public domain, so that others may freely build upon, enhance and reuse the works for any purposes without restriction under copyright or database law.
So yes, IANAL but AFAIK CC0 is just "public domain" but spelled out (the rest of that link explains why this might be needed).
> and thus an American can just assign it the 0BSD?
The catch, as I understand it, is that yes any American could take something under CC0 and re-release it under 0BSD (probably, again IANAL) but they can only give away the rights they had in the first place so they can't say "and nobody can claim a patent infringement on this", only "and I promise you can use this and I won't sue you for patent infringement". The original authors, if they have a patent, could (as this argument goes) still sue for patent infringement because they never agreed not to (allegedly).
1. Quietly patent sexy new algorithm
2. Release CC0 implementation of algorithm without mentioning they hold a patent.
3. ? (sue for infringement)
4. Profit
Anyone that tried this should be laughed out of court. If they're not, the law needs an overhaul.
How many times have we thought that in the last 10 years?
Corporations with lawyers up their ass aren't going to like something ostensibly put into the public domain by a German who technically can't do that, but lawyers don't like the WTFPL either, so it makes little difference.
> "hey use my stuff it's public domain"
> "ok thanks"
> "Ha! gotcha! Here's your lawsuit, pay up!"
For common people? Not a problem. No German has ever been sued for using SQLite.
> Every FOSS license must be interpreted as strictly as possible in case the copyright holder is bought by Oracle.
[1] https://gavinhoward.com/2022/01/the-law-of-strict-licenses/
If you're a billion dollar corporation, then you've already lawyered up and prepared for innumerable scenarios. If you're Joe Random like most of us, just get over it and move on with your life.
It's legal in America, but you'll have trouble in a lot of other countries (Germany for one).
It’s what I use for a public-domain-equivalent.
For reference, the OSI's comments when approving The Unlicense:
> There is general agreement that the document is poorly drafted. It is > an attempt to dedicate a work to the public domain (which, taken
> alone, would not be approved as an open source license) but it also
> has wording commonly used for license grants.There was some discussion
> about the legal effectiveness of the document, in particular how it
> would operate in a jurisdiction where one cannot dedicate a work to
> the public domain. The lawyers who opined on the issue, both US and
> non-US, agreed that the document would most likely be interpreted as a
> license and that the license met the OSD. It is therefore recommended
> for approval.
=> https://lists.opensource.org/pipermail/license-review_lists....
You seem to interpret "the document would most likely be interpreted as a license" as standing alone, which forms the basis of your critique. I interpret that line in the context of the sentence before. That is to say, I think "the issue" is how the license would be treated in a jurisdiction without a public domain, and I think that the lawyers agree the that it would be treated as a license only in that context.
To quickly address some other points:
Many people comment that "public domain dedication" is an ill-defined legal term and you shouldn't even try. Okay, fine, whatever. I'll happily use a public-domain-equivalent license. Citing the above, The Unlicense works for this purpose.
Your comments on its name are understandable, but as someone who is trying to actually license my code in a public-domain-equivalent way, I can't afford to be picky about that.
I don't trust/understand The FSF's "recommendation" for CC0 because in their guidance on CC0, they explicitly do not recommend using CC0 for software. (Not even to mention the thread we're commenting on.)
Many people on the OSI mailing list harshly criticized The Unlicense for being written by someone that isn't a lawyer. While I understand their concerns, again, I'm not aware of any licenses written by lawyers to be public-domain equivalent. The BSD 0-clause is just deleting portions of the BSD license. Since neither of these were written by lawyers, I feel like I can speculate by saying that The Unlicense seems to me to cover a lot more cases and is a lot more explicit about its intent than 0BSD.
All that is to say, I know that a lot of people don't like The Unlicense. I would love to use a public-domain-equivalent license written by a lawyer with a patent disclaimer. But doesn't exist, and the closest thing in intent is The Unlicense.
As for Unlicense versus CC0 or 0BSD, and the patent law matters in question here: the Unlicense explicitly says it’s talking about copyright law, and limits the effects of its fallback license to that. And as a public domain dedication, it can’t exceed that scope. Thus, worldwide it will exclude any patent law considerations just as thoroughly and reliably as CC0 which explicitly reminds you of this fact. This is a decent chunk of the problems with both Unlicense and CC0: public domain dedications only cover copyright law, and consensus has been building that that’s not enough for open source.
If you want a public domain dedication: use CC0, because it’s fairly clearly uniformly better than Unlicense.
If you want a public-domain-equivalent license: use 0BSD or MIT-0, but certainly not Unlicense, because it’s a PDD with fallback license, not a PDEL, which means that it cannot possibly have any implicit patent or trademark grant; you could staple a patent grant on top of it (or better yet, in my opinion, just state clearly that there are no applicable patent rights).
The problem with Unlicense is that by bad drafting it exists partway between PDD and PDEL, and doesn’t do the best job of either. So choose which you want, and take the better option.
It would be nice to have a public-domain-equivalent form license that included a patent grant or release. But Unlicense is not close to that in intent.
It seems to me very clear that the only fully consistent position for Fedora here would be to disallow all forms of public domain dedication (most germanely Unlicense), unless accompanied by some supplemental remark on patent law.
The only issue that has cropped up from this is that Google employees can use but not contribute to these projects.
I also personally love the name of the Unlicense. Personally, I use it because I'm interested in a bit of advocacy against monopoly copyright interest.
To address your indictment head-on: you suggesting the 0BSD as a better alternative is really missing my point. The 0BSD is not an alternative for my use case. The Unlicense is one of the very few overt "political" acts that I inject into the software I produce. Its purpose is to make a statement. The 0BSD doesn't do that IMO, so it's not actually an alternative that meets my advocacy goal.
You and Rick Moen seem to have the same apparent blind spot for this. See my conversation with him that started here (which might also clarify some aspects of my own position): https://github.com/docopt/docopt.rs/issues/1#issuecomment-42...
And finally, note that my dual licensing scheme is exactly a response to the "problems pointed out by quite a few people": https://github.com/BurntSushi/byteorder/issues/26
In suggesting 0BSD, I am indeed talking on legal grounds. For Unlicense, the only part of it that would seem to be any kind of “statement” is its name, which… well, you know my opinion of the misbegotten name. I feel like there must surely be some better way of making that statement (which I agree with!), but I don’t know what.
(I’ve been having fun using BlueOak-1.0.0, which is generally superior to the likes of MIT and Apache-2.0, and have been rather tempted to release things under only that license, and tough luck to anyone that cares about OSI approval, as BlueOak-1.0.0 is not OSI-approved largely because the authors are unimpressed with OSI process after having been involved with it for years in the past, but so far I’ve been doing silly things like trial-licensing BlueOak-1.0.0/MIT/Apache-2.0 to satisfy people’s sensibilities. I’ve also been wondering if I should just throw in the towel and go 0BSD, not like I care about the attribution anyway, or would ever pursue any license violation of any form for reasons of conscience. I had thought of going CC0-1.0 or similar, but legal limitations of PDD versus licenses decided me against it when I was mulling it over. Pity public domain dedication is so hard to do properly and so surprisingly limited in scope.)
> The Unlicense is a template for *disclaiming copyright monopoly interest* in software you've written
But yes, the name is a big part of the "statement," I agree.
> Pity public domain dedication is so hard to do properly and so surprisingly limited in scope.
Indeed. That should change.
> Part of the background story is that on patents, we've fought the scenario that a patent holder could offer some FOSS library implementing their patented method, and then come chase you for patent infringement. What are the boundaries for me, FOSS user, to not be ensnared by a submarine license?
Honest question: How could they claim damages from someone to whom they've granted permission? It just doesn't seem like a likely problem. So i'm trying to understand what the actual threat is.
That's actually the underlying rationale here, although it isn't explained well in the linked page. There is the possibility of an implied patent license grant (or waiver) from the licensor in a license that is silent about patent rights, for exactly the reason you state.
But CC0 is different, in that it grants permission to use while simultaneously being explicit that there is no implied patent license (or waiver) granted.
I think that's the key: technically they haven't granted permission, not explicitly, absent a patent waiver. (Disclaimer: not a lawyer, barely paying attention to this.)
Any lawyer trying to sneak out of patent claims by saying "oh but the FOSS license didn't state they were retaining patent rights..." is just regular legal bullshitting and would not stand long in any half-working justice system.
If a license has never been considered an implicit waiver, it seems overkill to say that "a license that says nothing about rights is materially better than a license that explicitly does not waive rights," no?
Yes, and I suspect a lawyer for a sufficiently-malicious GPLv2 dev looking to submarine an org could probably even take it further – because the FSF felt it was necessary to create GPLv3 in order to explicitly grant patent rights to GPLv3 licensors of the code, choosing to continue to use GPLv2 like my client the malicious-submarine-dev did, instead of relicensing in GPLv3, was done precisely in order to signal that they are expressly choosing to grant a copyright license but not a patent license to users.
Singling out CC0 seems arbitrary – if Fedora wants patent certainty, they should be sticking to things like GPLv3 and Apache
If you choose a license which explicitly says you're not waiving your patents, it insinuates that you are holding some, and they are used in the project.
Not anymore than the BSD-no-nuclear clause insinuates that there's a version somewhere that is suitable for being run in the control system of nuclear power plants, it doesn't.
CC0 just makes it clear what is in-scope for the license – copyright, not patents, because it's intended as a formalization of public domain and that is not something that can be formalized re: patents – unlike BSD or GPLv2, which implicitly are about copyright but never actually come out and say so (and if you were looking to screw someone over with some patent you hold, you're better off with something that doesn't even mention it, like GPLv2, rather than prompting the lawyers to explicitly ask about patents because the license reminds them that they're not covered)
The fallback license in the CC0 is supposed to be just a begrudging inclusion that applies in jurisdictions where the assignment to the public domain isn't upheld; the author who releases a work under the CC0 would be most pleased if there were not such jurisdictions, and the public domain assignment were enough.
OK, so then you have this fallback license which says, "I am not granting any use of the patents covering the ideas in this work". That's completely stupid; nobody who ideologically favors the public domain would want to use this.
The only rational use for CC0 is by a corporation which has developed some patented technology (e.g. video encoder) and wants to promote a reference implementation, to be used by paying patent licensees for making applications (plus individuals for educational, non-commercial, personal use).
If a fallback license of a public domain assignment says anything other than "you can use and redistribute this as you wish", it is not an earnest, genuine, ideological public domain assignment. Someone who wants to make such a grant to the public, yet uses the CC0, has been duped.
Probably the reasoning is that most licenses that don't mention patents don't actually have any patents behind them.
Most CC0 probably don't either, but since CC0 explicitly says patents are not waived, that's where Fedora decided to start cleaning house.
As a free software developer, I just want to know whether this is the way the wind is blowing? Do I have to re-license everything with a different BSD license that incorporates a patent waiver?
I see this whole thing as pointless, because even if you don't hold any patents, it's almost impossible to write any significant piece of code without infringing on multiple ridiculous software patents that someone filed somewhere.
Just because I add a sentence saying I won't enforce any patents against the users of my code doesn't mean that some third party in the world will not do that. That's probably one reason licenses are mum about patents; patents are a threat coming from someone else which your own waiver doesn't dispel.
The idea is that a generic grant of usage rights will cover multiple forms of IP, so it's fine.
But if it says "this is copyright only" that's a problem.
The issue isn't directly about whether it mentions patents or not.
Its the opposite.