![[RSS]](/static/img/feed-icon-14x14.2168a573d0d4.png){kind=icon}
Conservancy BlogDisplaying posts
by Bradley M. Kühn
by on April 16, 2026
This article discusses a current-headlines situation regarding Affero General Public License, version 3, Section 7, paragraph 4 (AGPLv3§7¶4.). I begin however with an explanation of the problem that clause sought to solve and how the clause works. This may seem an estoric license issue, but in fact this issue regularly impacts users today — particularly with the advent of “badgeware” (software that allows redistribution but includes annoying advertising that cannot be removed). Hopefully, this explanation helps readers understand the importance of the issue and gain vigilance when reviewing potential “further restrictions” placed on their copylefted software.
I began my work in copyleft licensing and policy in the late 1990s. In those days, there was a growing problem regarding usage of the GNU General Public License, version 2 (GPLv2) that threatened the software freedom and rights of users. It's a nefarious licensing slight-of-hand that works as follows:
The vendor seems to offer the software under a copyleft license. There's
a copy of GPLv2
in the top-level directory of the source code in a file called
GPLv2. All seems in order, and folks excitedly
engage in their right to copy, modify, and install modified versions of the
software. Maybe a few even think of a viable business idea that would
include (usually permissible) commercial redistribution of the software
for profit.
Unfortunately, someone notices a file
called LICENSE in the top-level directory that says:
Copyright (©) 1999, Sneaky Company, Inc.
This software is licensed under GPLv2, except that commercial modification and redistribution is strictly prohibited.
They've sadly discovered a self-contradictory license. Unfortunately, under GPLv2, these users are basically stuck; they have to go with the strictest possible interpretation given the self-contradiction. In essence, the licensor giveth, but the licensor immediately taketh away. In those days, these users couldn't start their business; they'd have to find or write another codebase.
I was tangentially involved with the drafting of GPLv3. On my list of issues to raise with the drafters was this very issue. By the time of GPLv3 drafting (circa 2006), this problem was rampant. Users were quite confused when they saw these self-contradictory licenses.
The solution was not obvious. Both GPLv2 and the earliest drafts GPLv3-family of licenses
(which includes GPLv3,
AGPLv3, and
LGPLv3)
already had this clause: You may not impose any further
restrictions on the recipients' exercise of the rights granted herein
(quoting the GPLv2 version; the
GPLv3/AGPLv3 version varies slightly).
The problem: this only prevented downstream licensors from imposing further restrictions. If a sole entity is the original author, copyright holder and initial licensor of the work — and if they hold those powers exclusively — typically that entity may issue a self-contradictory (or even completely incoherent) license. In that case, downstream licensees are awash with legal uncertainty. So, that “may not impose” clause just does not solve this particular problem.
A new clause was needed. Later drafts of the GPLv3 (which is almost textually identically to AGPLv3 — only §13 differ between the two licenses) included a fascinating solution in §7¶4:
If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term.
Copyleft is indeed most inventive when it empowers the downstream user (who, of course, is often just the next entity in a long distribution chain of (both commercial and noncommercial) software sharing). AGPLv3§7¶4 is an innovative and very necessary clause that liberates users who face the situation described at the start of this article. Had our excited new business seen this …
Copyright (©) 2008, Sneaky Company, Inc.… they could simply toss away the additional restriction like so …
This software is licensed under AGPLv3, except that commercial modification and redistribution is strictly prohibited.
Copyright (©) 2008, Sneaky Company, Inc.… and copy, modify, redistribute, redistribute modified versions and/or install modified versions of the software freely under AGPLv3's pure terms.
This software is licensed under AGPLv3, except that commercial modification and redistribution is strictly prohibited.
There remains one drawback to this solution: it demands courage from the user that strikes the “Further Restriction”. In theory, all is well and safe. In practice, the types of companies that pursue tactics like self-contradictory licenses and “gotcha” further restrictions are also the most predatory, unfriendly, litigious, and aggressive businesses. One who exercises their clear and correct rights under AGPLv3 will certainly face public condemnation, and possibly frivolous litigation.
For years, the Neo4j case was the primary exemplar of this phenomenon. SFC followed the case closely, I served as an expert witness for the Defendant, and SFC filed an amicus brief on the appeal. The lower court decision was highly problematic, and the case concluded with a voluntary dismissal (and as such has not been heard by any Appeals court). Thus, while the lower court decision does not create bad precedent, it also does not offer any good precedent for those corageous users who exercise this particular right.
(Full disclosure: SFC runs a self-hosted Nextcloud instance — but other than being fans of their work and satisfied users — we have no formal relationship with Nextcloud (or IONOS). We also certainly have no relationship whatsoever with Onlyoffice or Ascensio System SIA.)
A few weeks ago, yet another saga in the history of AGPLv3§7¶4 began. Ascensio System SIA published Onlyoffice with the sneakiest “Further Restriction” that I've ever seen. Ascensio's further restriction states:
Pursuant to Section 7(b) of the License you must retain the original Product logo when distributing the program.
Pursuant to Section 7(e) we decline to grant you any rights under trademark law for use of our trademarks.
This restriction is particularly nefarious because it is dressed in the trappings of explicitly permissible requirements in AGPLv3§7.
In addition to our beloved AGPLv3§7¶4, the rest of AGPLv3§7 includes some provisions designed for cross-FOSS-license compatibility. During the GPLv3 drafting process, a survey was conducted of all popular FOSS licenses. In the spirit of making sure no terms of GPLv3 inadvertently contradict a requirement in one of those licenses, AGPLv3§7(a-f) were devised for maximal cross-FOSS-license compatibility.
For example, the 3-Clause-BSD license's third clause states:
3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission
Now, get ready for Onlyoffice's trick, which takes three moves:
The trick is “too clever by half”. Ascensio have indeed created a self-contradictory license, since the users are prohibited from displaying the trademarked logo, yet the users are also forbidden to remove the code that displays that trademark logo.
We thought through this problem already years ago during AGPLv3's drafting. In fact, I recall much discussion to verify there were no strange interactions between 7(b) and 7(e) — and there are not. AGPLv3§7(b,e) states (including relevant definitions from earlier sections):
An interactive user interface displays “Appropriate Legal Notices” to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion. …
Notwithstanding any other provision of this License, for material you add to a covered work, you may (if authorized by the copyright holders of that material) supplement the terms of this License with terms:…
[7]b) Requiring preservation of specified reasonable legal notices or author attributions in that material or in the Appropriate Legal Notices displayed by works containing it; or …
[7]e) Declining to grant rights under trademark law for use of some trade names, trademarks, or service marks;
Nota bene: there is a very short list of things that a licensor can require
downstream licensees preserve. Logos, advertising, and similar are not on the list. Badgeware
is not a reasonable legal notice
.
Nextcloud has partnered with IONOS to fork Onlyoffice to create Euro-Office. To my knowledge, this codebase complies completely with the valid AGPLv3§7(e) requirement: Ascensio's trademarks have been scrubbed from the sources. However, since what Ascensio purports as a license term supplement pursant to AGPLv3§7(b) is actually a backdoor Further Restriction, Nextcloud and IONOS are completely within their rights (pursuant to AGPLv3§7¶4) to remove that further restriction.
As expected, Ascensio published an unfair, inaccurate, aggressive, and community-unfriendly response. I predicted long ago that adjudication of AGPLv3§7¶4 would require nerves of steel. This unwarranted denigration of an upstanding member of commercial FOSS community — Nextcloud — demands condemnation and Nextcloud deserves our thanks for confronting this bully.
As I did in the Neo4j case, I've told Nextcloud that I'm available as an expert witness if they end up in litigation with Ascensio. I hope, however, that Ascensio will realize their error, remove the further restriction themselves, and embrace the value of AGPLv3. Pure AGPLv3 was designed to faciliate web service business models. Our community would quickly and gladly embrace Onlyoffice, help its upstream development, and welcome them to the FOSS community. ∎
About the author: Bradley M. Kühn is a lifelong FOSS activist and has focused his career on FOSS licensing generally, and copyleft licensing in particular. Kühn invented the Affero clause in the original AGPLv1, and co-drafted AGPLv3§13.
Bradley is not a lawyer, SFC is not a law firm, and this article is not legal advice.
by on March 4, 2026
Earlier this week1, the U.S. Supreme Court (SCOTUS) denied certiorari (cert) in Thaler v. Perlmutter. Thaler contended that an image — generated by a Large Language Model (LLM)-backed Artificial Intelligence (AI) — deserved copyright registration. Since the U.S. Copyright Office refused to grant the registration, Thaler appealed to the U.S. District Court for the District of Columbia (DC Circuit). That Court affirmed the Copyright Office's decision. SCOTUS' denial of “cert” means they will not hear the case. Strictly speaking, this denial does not affirm the DC Circuit Court's ruling, but it does mean the DC Circuit decision stands.
Many in the Free and Open Source Software (FOSS) community raised concerns about the impact on copyleft — and even FOSS in general. TL;DR: Don't Panic! — this case is extremely limited in scope.
First, a proviso: this case is about copyright of an artistic image, not software. Copyright law — and the legal precedents around it — differ widely for different types of creative works. Analysis of the copyrightability of works of software varies in notable ways. Therefore, do not to assume that analysis for images apply broadly to software.
Second, while the decision is “published” 2, there are also many other cases related to LLMs and AI currently pending throughout the U.S. Courts. Courts and laws always lag behind technological advancement. Indeed, this is precisely why copyleft was invented: as a mechanism to achieve with existing laws and precedents what we could not accomplish in the legislature. Forty-one years after copyleft's invention, we still do not have a federal law that mandates software right to repair!
Third, the Court found that a registration was not valid (at this time) if the work's sole author is a computer program. Thaler (who was both (one of) the author(s) of that computer program and its user) repeatedly waived any claim to consider Thaler's own copyright in the LLM-backed AI prompting process. Thaler also did not argue any copyright interest in the LLM-backed AI system itself were subject of the registration. So, this decision does not evaluate any creative expression by (a) the author(s) of the prompts themselves, (b) copyrights held in the LLM, its weights, generation, curation, or its user interface, and (c) copyrights held in underlying works in the LLM training data.
Thaler's original registration was the root cause of this substantial narrowing because the registration contended that the AI system itself was the author of the image. This case only considers a copyright registration where the sole “author” is identified as a specific computer program. Thaler stipulated that the work was generated solely through prompts and no human modified the work thereafter. As such, even if the other districts begin citing this case regularly, and even if many districts decide it applies to software without further consideration of the difference in the types of works, such precedent causes no disaster for FOSS.
Admittedly, some LLM-backed generative AI agents can be merely prompted to create a work of software from scratch that has some transient utility. However, the most common workflow in using these agents (at least in FOSS development) is as follows:
Furthermore, dicta 3 — appearing in the DC Circuit ruling — supports a conclusion that the human actions on that third step would constitute creative expression — affixed in tangible medium — suitable for copyright registration. Indeed, their ruling states:
First, the human authorship requirement does not prohibit copyrighting work that was made by or with the assistance of artificial intelligence. The rule requires only that the author of that work be a human being — the person who created, operated, or used artificial intelligence — and not the machine itself. — (Thaler v. Perlmutter, 130 F.4ᵗʰ 1039, 1049 (D.C. Cir. 2025))
The Court also indicated these other issues are for a future time in another case. The DC Circuit readily admits that their ruling applies only to the state of AI systems at the time of writing4. Again quoting from their ruling:
Of course, the [Thaler's AI] Machine does not represent the limits of human technical ingenuity when it comes to artificial intelligence. Humans at some point might produce creative non–humans … Science fiction is replete with examples of creative machines that far exceed the capacities of current generative artificial intelligence. For example, Star Trek’s Data might be worse than ChatGPT at writing poetry, but Data's intelligence is comparable to that of a human being. See Star Trek: The Next Generation: “Schism” (Paramount television broadcast Oct. 19, 1992) (“Felis catus is your taxonomic nomenclature, an endothermic quadruped, carnivorous by nature”). There will be time enough for Congress and the Copyright Office to tackle those issues when they arise. — (Thaler, 130 F.4ᵗʰ at 1050)
[ As always, SFC is not a law firm, IANAL, and TINLA. ]
1 I — and my colleagues at SFC — acknowledge that SCOTUS made other decisions recently regarding an array of important social justice causes. Since SCOTUS' decision to deny cert in this particular case is so closely related to my work, I'm writing about it. However, all of us at SFC acknowledge that our community is reeling from other recent decisions.
2 In this context, “published” is a term of art that lawyers use to describe a case that the publishing Court (in this case, the DC Circuit) felt was important enough to “share officially and formally” with other Courts. While (in a precedent-based legal system) any Court can cite an unpublished case from another Court, published cases are much more likely to be cited than unpublished ones.
3 “Dicta” is explanatory language found in a court's decision that isn't necessary to the court's conclusion. . Dicta isn't precedential but it can be persuasive.
4 Note that the DC Circuit issued their ruling in March 2025. It is not uncommon for SCOTUS to delay for a year (or more) before issuing a ruling to grant or deny cert.
by on January 26, 2026
We at Software Freedom Conservancy are disappointed at some surprising news. Two weeks ago (THU 2026-01-08), we had our original pretrial motions hearing scheduled in our historic impact litigation against Vizio. Just about an hour before the hearing's start-time, Judge Sandy Leal issued a minute order that rescheduled the hearing and (effectively) removed the trial (which was set to start on Monday 12 January 2025) from her calendar.
The rescheduled hearing date was Monday 2026-01-26 at 09:00. At 08:15 that morning, our attorneys were contacted from the Court Clerk that the hearing was again postponed..
We have been in this litigation against Vizio since October 2021. Vizio violated both the General Public License (GPL) and Lesser GPL Agreements. Vizio's “Smart” TV products include more than a dozen packages under these copyleft licenses, yet Vizio has continually failed to comply with these agreements in various ways — most notably (and including but not limited to) by (a) not providing complete, corresponding source code, (b) not providing “the scripts used to control compilation and installation of the executable[s]”, and (c) not providing object code necessary for relinking the LGPLv2.1'd works. We were looking forward to our days in Court that week to show the world all the details of Vizio's non-compliance, and to ask the Court to acknowledge (among other things) our right as a third-party beneficiary under the GPL Agreements to receive all the materials that those Agreements require Vizio to give to all consumers who purchase their devices. These devices, BTW, are called “Smart” TVs because what's inside is actually a small (but powerful) computer attached to the giant video display — driven and controlled largely by copylefted FOSS.
Notwithstanding our frustration, our trial was delayed for good reason. Another case — even older than ours — needed more time for their jury trial (and thus had priority over ours). While some criticize the USA for being “too litigious”, we at SFC believe firmly that the civil Courts are the best place where ordinary citizens and small, scrappy non-profit charities like SFC can seek justice when our rights are violated. We also know that there is more injustice in our country these days than anyone would like, and this delay occurred because there are other folks out there seeking justice on other important issues and rights, too.
We understand that we've been waiting for a long time in a very long queue in the California Courts, and while we (like everyone) get frustrated when the line is taking much longer than expected, we also appreciate that Judge Leal is carefully managing her docket to grant all parties an impartial opportunity for justice.
Attorneys for both SFC and Vizio are now negotiating with the Court for rescheduling. We hope the pretrial hearing will be scheduled fairly soon. We will update here and on the Fediverse as we know more.
We'll spend the next few weeks posting the various recent motions and filings in the case, and publishing some retrospective summaries of the last four and a half years of the case for you all to read.
Be sure subscribe to our feed in your RSS readers/aggregators and follow us on the Fediverse (via Mastodon or your preferred ActivityPub software). to receive updates!
by on December 23, 2025
I wrote last month about my diabetes diagnosis this year and my difficult choice to wear a proprietary device (called a CGM) on my arm 24/7 to continuously monitor my glucose levels. Like my friend and colleague, Karen M. Sandler — who previously made a much higher-stakes choice to receive a proprietary implanted defibrillator to keep her safe given her genetic heart condition — I reluctantly chose to attach proprietary hardware and software to my body.
The device itself is quite proprietary, but fortunately the FOSS community has reverse engineered its activation and data collection protocols — creating an Android application that does a better job than the manufacturers' proprietary ones0.
Here in the USA, we strangely use capitalism as the center of our health care system. Two major for-profit competing brands of CGM are available here. My diabetes specialist prefers the (ironically named) Freestyle Libre Plus from Abbott. I (also rather strangely) bring a prescription for electronics to a pharmacy every month. On 2025-12-03, that pharmacy sent me an alarming text message (shown here).
After reading that text, I found the USA FDA announcement. My spouse cross-referenced the lot numbers while I read them off from all my Freestyle boxes1. I had indeed recently worn an impacted device!
Only because my diabetes is so early of a stage was I relatively safe. The FDA reports that Freestyle injured over 700 people and killed seven people with this bug. Specifically, the bug caused the device to falsely report an extremely low glucose level. Advanced stage diabetics use low reading information to inform them that they may have too much insulin currently. The usual remedy is to eat something sugary to raise glucose in the blood. Such should be done only with great care, as a false low reading can harm and even kill the patient (who eats a high-sugar-content item while glucose in the blood is, in fact, not low).
Proprietary software in medical devices harming patients is not new. In 1985, the Therac-25 killed three people. In 2020, hundreds of patients who relied on a financially troubled tech startup found their occular implants suddenly unsupported. Some patients went blind as the devices powered down without updates. There are more examples that I could include here, but rereading these horrific stories is frankly more than I can take right now when I think of fellow diabetes sufferers who were “killed by code” recently..
It's hubris for activists to guarantee that harm would be prevented if Freestyle had publicly released the hardware specifications and the complete, corresponding source code (CCS). FOSS isn't immune to bugs — even dangerous ones. However, in the centuries since the Enlightenment, we know that the scientific method depends on public disclosure about data and wide-reaching peer review of past work. FOSS (plus a publicly disclosed hardware design) wouid allow the millions of hardware and software engineers to peer-review the integrity, security, and safety of the devices to which patients entrust their lives. We achieve the promise of humanity when we each entrust our safety and health to our entire community — not merely a single for-profit entity.
We also will probably never know whether this issue was in hardware or software. The bug disclosure is incredibly vague, and it remains unclear how much investigation was done (if any) by government regulators into this problem. As a public policy and public health matter, the public deserves to know the technical details (software and hardware) of both the functioning device and the failed devices. NGOs should be permitted to perform their own investigations and confirmations of public safety.
Given that the hardware, software, and medical for-profit industries refuse to put the rights, safety and security of patients first, wrongful death lawsuits are typically the only way to hold these companies accountable. Yet, there are very few people who have not agreed Abbott's toxic terms of their proprietary companion application — I guestimate that fewer than 1% of Freestyle-using patients have used Juggluco from their very start (and thus never agreed to Abbott's terms). This is significant because Abbott includes a comprehensive one-way indemnity for themselves in the terms. I hope that a class action suit begins soon on this matter, but I wonder and worry that so much of the class may have signed this indemnity (which may make the road to justice bumpier).
Finally, I want to offer that if there is anyone out there who does tear-downs of extremely tiny electronic devices, I would be thrilled to find a volunteer who would like to see if we can either extract any software components from the device, or reverse-engineer the hardware. I have saved and sanitized all of my prior CGMs. I'd gladly send one along to anyone who wants to give a try at taking them apart. (Contact SFC or contact me on the Fediverse (via Mastodon) if you're available to do this work.)
For my part, I look forward (after the Vizio trial) to sending some patches to Juggluco and also getting Juggluco available in F-Droid. Our best option in the face of these powerful medical device companies curtailing our rights is to invest our volunteer time into the edges where FOSS has resiliently worked around the constant roadblocks erected by bad actors.
0My prior post about CGMs discussed the GPLv3'd Juggluco in more detail.
1 In a fascinating turn of events, at least one of my past monitors (of which I fortitously saved all the boxes with the lot/serial number on them) is listed in the FDA's spreadsheet as recalled lot, yet the serial number is listed as “ safe to use” on Abbott's webform 🤔 … I'm left wondering how I can trust Abbott to write reliable software stuck into my arm if they can't even write a web form that cross-references serial numbers to lots correctly 😬.