-*- mode: outline -*- * The Public Domain Problem The public domain is a marvellous thing. Untold cultural riches freely available for anyone to use! What could be nicer? ‘Public domain’ describes works which are not protected by copyright. There are two ways that things end up with this status: • By age. Usually, once the original author has been dead for 70 years (although the exact amount of time varies worldwide), everything they made is no longer protected by copyright. Every year on 1 January, the works of all the authors who died the previous year minus 70 become freely available for all to use. • By dedication. Someone who owns copyright in something can write an explicit note saying they don’t want the work to be protected by copyright any more. Then anyone can use the work without restriction, even if all the other works by the same author are still protected. For artists and software developers who want people to enjoy their work as much as possible and don’t care about remuneration or even credit for their work, a public domain dedication is an attractive way of spreading their works as widely as possible. Since copyright is given to authors automatically without them having to register it, apply for it, or otherwise explicitly claim it, an explicit dedication is needed for something to be in the public domain this way. I’m going to divide public domain dedications into two kinds: informal ones, and Creative Commons Zero. (There could be other non-informal dedications in the latter category, but I’m not currently aware of any.) I’m also mostly going to talk about copyright in software (although to some extent to other kinds of copyrightable works). Informal public domain dedications for software often take the form of comments in code which might look something like this (an anonymous selection from GitHub Code Search results): > This source code is offered for use in the public domain. You may > use, modify or distribute it freely. > This program is granted to the public domain. > These files are all under public domain. > This is free and unencumbered software released into the public > domain. However, as we’ll see, there may be reasons to doubt whether these little notes are effective, which is why using Creative Commons Zero is a good idea. But certain organizations have conjured up problems with Creative Commons Zero as well! Together, these issues rise to The Public Domain Problem for software authors. ** The legal situation Very few countries – none that I consider directly relevant to me – have an explicit statutory provision allowing copyrights to be abandoned to the public domain by a deliberate act of their controller. In the United States, the situation at common law is as clear as it can be short of a Supreme Court ruling. The Ninth Circuit appeals courts belong to the next level below the Supreme Court, and they have decided multiple times that public domain declarations are legally effective in permanently abandoning all copyright. Precedents like this are legally binding in the United States in the same way statute law is. In Germany, as Daniel J. Bernstein[*] points out, there is one section in the Copyright Law (§ 33 UrhG) which considers what happens in the situation when ‘the rights-holder renounces his right’. But if one interprets that section in its entirety, it seems to contradict itself if one assumes that it is talking about the public domain: it says that grants of usage rights, including grants of *exclusive* usage rights (which is the German legal equivalent to sale of copyright), remain in effect when the owner of the rights changes or when they renounce their rights. So an exclusive grant of rights can remain exclusive even if you renounce your copyright, which does not make it sound like a copyright abandonment is actually effective. It *definitely* doesn’t sound like something that is possible in the U.S. – you transfer ownership of your copyright (= give exclusive usage rights) to someone else, maybe your employer, and they later release it to the public domain – is allowable in Germany. [*] https://cr.yp.to/publicdomain.html Bernstein points to several court decisions in Germany which have recognized public domain dedications, which is encouraging. However, Germany does *not* work on a precedent system – only statute law is binding, and each court can re-interpret statutes as it sees fit in each individual case. All it takes is someone taking a case about the validity of the public domain to the Landgericht Hamburg (every copyright troll’s favourite European court) and it could be decided very differently. Bernstein also points to an informational page by the Intellectual Property Helpdesk of the European Union which states that works can be in the public domain by dedication in Europe. I contacted the Helpdesk to ask what their statutory or judicial basis for this claim is, and they admitted that there was none, and that it is down to the laws of individual EU countries. This is the first part of the Public Domain Problem: while in some countries it’s fairly clear that you can abandon your copyright, in others it’s not clear. ** Why’s there a problem anyway? ‘So what’, you say. ‘I’ve put the public domain notice on my work. I’m promising I’m not going to threaten or sue anyone for using my work. Shouldn’t that be enough?’ Unfortunately, you are not the only person who will exercise control over your intellectual property. At some point, you will die, and the control will pass into the hands of your heirs and successors. (One of the absurdities of the 70 years p.m.a. copyright term is that every copyrightable work created by an adult will almost invariably be controlled by others after their death for longer than it was controlled by them personally while they were alive.) Ultimately, you have no idea who will end up in charge of your copyrights. If you create a valuable work and declare it to be in the public domain, your legatee may decide to try to challenge whether it’s actually possible to do that with legal effect in whatever jurisdiction they think they can try it in. This creates unnecessary risk for the users of works which are in the public domain by dedication, rather than by age. Also, the history books of free software and free content are not exactly devoid of authors placing their software under a software licence which allows unrestricted use, who later start threatening someone who uses their stuff in a way they don’t like. (One such case is going on even as I write this essay.) So users are justified in wanting the peace of mind of knowing that the legal instrument they received your software under is actually legally effective. Permissive licences explicitly give the recipient the right to use, modify, and distribute the software in any form whatsoever. Examples of such licences are the MIT and BSD licences for software, and the CC-BY licence for other free content. However, they do still impose restrictions on redistribution: you have to tell everyone who owned the copyright at the time when you got the work, and give them the original licence as well. (The original licence is not necessarily binding on people who get (your modified version of) the software from you, unlike with non-permissive/copyleft licences. But you do have to tell them what the original licence terms were.) The public domain has none of these restrictions, but, curiously, it seems to be harder to effectively remove these restrictions worldwide than it is *not* to remove them. ** Digression I: Other legal rights and responsibilities (You can skip this section at first and then refer to it later if some of what I say after it doesn’t immediately make sense to you.) *** Trademarks If you want to make something available under a particular name or logo, and you don’t want others to be able to compete with you using the same name and logo, you can get trademark protection. This is distinct from copyright in several ways, including because copyright does not generally recognize a single word name as being copyrightable. Trademarks have some key differences from copyright: • you have to apply for (and pay for) a trademark – you don’t get them automatically; • each individual trademark usually only applies in one country, so you have to apply for (and pay for) trademarks in each individual country where you want protection; • there is a centralized public register of all trademarks, so you can easily find out whether a name is already trademarked; • a trademark’s term is indefinite, rather than fixed – as long as the trademark’s owner is clearly still using it commercially, they can continue controlling the name; • a trademark is, to some extent, tied to the kind of product or service that is actually sold under it. The free software community has generally agreed that software being distributed under a trademarked name does not make the software non-free. There have been occasional spats about trademarks, but the software itself remains free even if names have sometimes had to be changed by distributors because of trademark restrictions. *** Patents Copyright applies to real creations, not to ideas. Patents can be taken out on ideas to stop others making and marketing products that use those ideas. Patents have some similarities with trademarks: • you have to apply for (and pay for) a patent; • each individual patent usually only applies in one country, so you have to apply for (and pay for) patents in each individual country where you want protection; • there is a centralized public register of all patents, although in some countries there is an annoying exception here (see below). And some differences that make them more like copyright: • a patent has a fixed term, usually of 20 years. In many countries, mathematical formulas and proofs are supposed to be unpatentable. Since a computer algorithm is just a series of formulas, and a computer program is mathematically equivalent to a proof, computer algorithms and programs are supposed to be unpatentable. Unfortunately, dysfunctional regulatory systems around the world have been bamboozled into allowing patents on software ideas anyway. This may happen infrequently or it may be considered quite normal, depending which country it is. In most countries, patents are published as soon as they are applied for, so everyone knows what might be subject to a patent claim in the future. In other countries, patents sometimes remain a secret until they are actually granted. This means that someone can apply for a patent, publish software which uses it, and then only once the patent is granted start to wring people out for patent licence fees, although nobody could have had any idea when they started using the software that it was covered by a patent. The free software community perceives a great threat from software patents. We’ll look at how they have responded to that threat below, but unlike trademarks, the consensus is that software which implements ideas that are known to be patented is not really free software. At most, it’s free software but with a dangerous asterisk attached to it. Despite this, most free software licences do not explicitly mention patents. *** Warranties Another thing that typical free software licences do is disclaim all warranties. Typically, the supplier of a product or service is required to vouch for it being of good quality. If it isn’t of good quality, they may have to fix it at no (extra) cost, or compensate the recipient for any damage caused by using their poor-quality wares. This is undesirable for authors and distributors of gratis software of any kind, since they didn’t get any money for making the software, but may incur costs in servicing the legal obligations of a supplier if the software turns out not to work. So besides granting copyright, a typical free software licence will disclaim all warranties, including the ones that the law usually gives to the recipient of goods and services automatically. (These are the bits of free software licences which are typically in ALL CAPS SHOUTING. The law of the United States requires certain clauses in agreements to be extra extra prominent to have any legal effect, including disclaimers of warranty. This law does not apply in other countries, so e.g. the EUPL, which is for European software authors, does include a disclaimer of warranty, but it doesn’t shout at you.) Warranties have nothing to do with intellectual property, and can be a matter of concern for a distributor as well as for the original author. For example, if Alyssa P. Hacker releases her software with a warranty disclaimer, but I don’t get it from her directly but from the AmazingFreeOS package distribution servers, neither Alyssa nor the AmazingFreeOS maintainers can be held responsible for anything that goes wrong. ** An attempt to solve the Public Domain Problem: Creative Commons Zero Recognizing that informal public domain dedications might be legally problematic in some jurisdictions worldwide, the Creative Commons organization has published the Creative Commons Zero (CC0) dedication,[*] intended to be a globally legally watertight way of releasing all restrictions on a copyrightable work. [*] https://creativecommons.org/publicdomain/zero/1.0/legalcode Creative Commons Zero takes a three-tiered approach, in order to provide recipients of CC0 free content with a maximal guarantee that the author or their successors will not try any take-backsies. In the first tier, it declares that all copyright in the work is abandoned. In case that isn’t effective in some jurisdictions for some rights, it then says that any rights which aren’t effectively abandoned by this are licenced unconditionally to everyone who wants to use the work. In case *that* isn’t legally effective, it then promises that the controller of any remaining rights in the work won’t sue anyone. There should be no better way to place something in the public domain than CC0. Unfortunately, paranoia in the free software community about patent rights has cast a black cloud over this superb legal tool. The last section of the CC0 code is entitled ‘Limitations and Disclaimers’ and, as you’d expect, is full of the usual sort of stuff that comes right at the end of legal documents: making explicit what should have been obvious, just in case. There’s also a warranty disclaimer here, but mostly it says stuff that should have been obvious: the Creative Commons organization wrote this legal text, but that doesn’t make them a party to it; if there are any other copyrights in the work, the person placing their work in the public domain with CC0 can’t affect those and it’s not their responsibility to get clearance for them. The same things apply implicitly to any other free software or free content licence: the Free Software Foundation is not automatically a party to the GPL as applied to any particular software; and as a basic matter of law, you can’t grant permissions that aren’t yours to grant. But there is also an explicit statement that CC0 only affects copyright, and not trademark nor (crucially) patent rights. Again, this should have been obvious to anyone reading the text so far: like a typical free content licence, it talks only about ‘copyright and related rights’ in the previous sections, and the definition of these makes clear that trademark and patent rights can’t be considered ‘related rights’. In particular, it defines ‘copyright and related rights’ as being granted by laws automatically, and not only upon application.[*] [*] An example of a non-copyright right which is still automatic is European database rights. Look it up if you care. CC0 is intended to cover more than just software – it’s designed for any kind of free content. In particular, someone in the physical sciences might want to put out an academic paper under CC0, and that paper might describe a patented idea. Neither that person nor their employer is likely to want to completely waive their rights to that patent just by virtue of having described it in a CC0-dedicated document, so they appreciate an explicit statement to the effect that only copyrights are affected. Unfortunately, this explicit statement that patent rights are not granted by CC0 has spooked some of the free software certification organizations. It began with the Free Software Foundation, who said that CC0 was okay for releasing free software, but users should be sure to get a patent grant alongside it.[*] Then the Open Source Initiative refused to certify it as an ‘Open Source licence’ – not because it doesn’t actually meet the requirements for capital-O capital-S ‘Open Source’ merely because it would allegedly set a bad precedent for them to begin certifying instruments which *explicitly don’t grant* patents, as opposed to the numerous licences they already certify which merely *don’t explicitly grant* patents.[†] Finally, in 2022 the Fedora distribution of the Linux operating system began refusing to redistribute software that carries a CC0 dedication due to this part of the text.[‡] [*] https://www.gnu.org/licenses/license-list.html#CC0 [†] https://opensource.org/faq#cc-zero [‡] https://lists.fedoraproject.org/archives/list/legal@lists.fedoraproject.org/thread/RRYM3CLYJYW64VSQIXY6IF3TCDZGS6LM/ But since this clause of CC0 is obviously merely clarifying what was already implicit above it – indeed, most other free software licences don’t explicitly mention patents either, and almost no informal public domain dedications do – what gives? ** Digression II: ‘Implicit patent grant’ theory Free software licences are usually understood as affecting copyright and warranties only, unless they explicitly mention other forms of protection. Some have argued that even those which don’t mention patents often implicitly give users and distributors of the software patent rights. For example, the MIT licence says that the recipient of the licence gets: > without limitation the rights to use, copy, modify, merge, publish, > distribute, sublicense, and/or sell copies of the Software which seems not only to affect copyright but potentially patent rights as well, since you couldn’t do those things with the software without limitation unless the licensor granted you the rights to use any patented ideas within the software. Implicit patent grant theory has never been tested in court anywhere in the world. It’s also not entirely clear (to me, at least) why the logical conclusion of accepting this theory would not also be an implicit trademark grant, at least inasmuch as the trademarked names are included with the software files (which is usual), but the proponents of this theory don’t seem to want to go that far. ** The Public Domain Problem in full Of all possible licensing arrangements, the public domain ought to offer the recipients of all kinds of works, including software, the strongest guarantee that they can do whatever they like with the work without restriction. Unfortunately, it is uncertain worldwide whether things which are claimed to be in the public domain, but which are still within what would otherwise be their normal term of copyright, are actually in the public domain. Creative Commons Zero should remove *this* uncertainty entirely, and indeed it does. For works which are software, though, some organizations claim that it is dangerous for users and distributors because of patent rights. Frustratingly, those same organizations are generally happy to accept not only works under licences with no explicit patent grant, but also accept works under informal public domain dedications which might not actually be effective in terms of copyright either. Informal public domain dedications also rarely release warranties, which exposes both authors and distributors to some risk, but the organizations which reject CC0 are likewise generally happy to accept works without a warranty disclaimer. ** Possible solutions *** ‘Pull your heads out of your arses!’ The FSF, OSI, and the Fedora Project could simply be convinced that their fearmongering is over nothing, and that it is not logically consistent to express concern over a licence which *explicitly does not* grant patents while saying nothing about licences which *do not explicitly* grant patents. Unfortunately, this is unlikely to happen. Indeed, as the Fedora Project’s comparatively recent rejection of CC0 for software shows, this paranoia is likely only to spread to other organizations over time. So something needs to be done. *** CC0 + a patent grant I have no patents to grant and do not want to give anyone the misleading impression that I do. However, the Free Software Foundation says that CC0 is entirely okay as long as it’s accompanied by a separate grant of patent rights. The OSI has never explicitly allowed this, though, so software distributed under such an arrangement would still not be ‘Open Source’ with capitals. Fedora has also not stated whether they consider this sufficient, so they (and any other free software distributor which takes a similar stance) might refuse to package your software even with a patent grant. Also, there are no real standard documents for patent grants in software that aren’t tied to free software licences. While licence proliferation might be less of a concern for patents than for copyright, free software communities generally take a dim view of home-grown legal documents accompanying software releases. *** Dual licensing If one believes in implicit patent grant theory, one could also choose to put software under both a CC0 public domain dedication and under a licence which is accepted by organizations such as FSF, OSI, and Fedora. Since the Creative Commons Zero dedication abandons copyright where possible, and ordinary licences such as MIT, BSD etc. *depend* on copyright to be effective, a person who releases their software under both a public domain dedication *and* a licence is inherently performing a contradictory act. Note that this is distinct from the second tier of the CC0 legal code, which explicitly states that it only applies if the first tier, the abandonment, is ineffective for some or all rights. Since patents are the only remaining concern for users and distributors after a CC0 dedication, a document which clearly does nothing more than grant patent rights ought to be sufficient, but we don’t know that the fearmongerers who reject CC0 all agree that it is. *** Revising CC0 CC0, with the explicit note that patent rights are not affected, is an attractive tool for non-software authors who do actually want to retain patent rights. So simply deleting this clause or adding an explicit patent grant in a revised version of CC0 would make it more useful for some and less useful for others. Instead, a future CC0 2.0 code could amend this clause so that it only applies if the work being released is not a work of software. Alternatively, an even freer version of CC0 could be released specialized for cases where authors *do* want to release patent (and possibly trademark) rights. This CC0+ (perhaps, in a nod to its intended use for software and the vagaries of the IEEE 754 system for arithmetic software, it could be called CC Negative Zero) would be almost identical to the existing CC0, but with the offending clause changed to an explicit patent grant. ** Conclusion I think the Creative Commons organization should do software authors a favour and give serious thought to removing this unnecessary dark cloud over CC0-dedicated software by offering an alternative to, or a revision of, the CC0 legal code. written by Daphne Preston-Kendal, upon Michaelmas in the year of our Lord 2024, Berlin, Germany. The holy angel protect me from anyone claiming me to be a lawyer, to be their lawyer, to be giving legal advice, or spreading other falsehoods about me or about this document.