How do trademarks apply to open-source?


In the early 1990s, there is a company operating under the name Portable Network Graphics, who brands their products with the mark PNG. They register these trademarks to protect their brand.

A few years later, some developer(s) release a specification and reference implementation for an image format called Portable Network Graphics, suggesting it be adopted with the abbreviation PNG. The image format specification is released to the public domain, and the code is released under an open-source license such as MIT.

This image standard is applicable to the industry which PNG (company) operates in, and now other actors in their industry are using this standard for various applications, commercial and otherwise. As the specification and code are open, various individuals and groups create forks that are hosted on multiple websites (GitHub, SourceForge, etc.)

Could there be a trademark infringement here? Who may be liable?

Would it make any difference if the developer(s) release the specification and code anonymously or under pseudonym?

Would it make any difference if the name and abbreviation aren't exactly the same, as they are here?

My question's curiosity isn't specifically about the names mentioned above, but about how open-source developers and those who use open-source software should think about trademarks (and whether it's any different than a "normal" business or product).

Since open-source is inherently not limited to one jurisdiction, feel free to answer with regard to any legal system(s) in particular.

I have started a wiki answer to collect the information and simplify it for developers. Please consider contributing to that, as it's just my interpretations as someone who doesn't study the law.

  • 2
    Without doing a full analysis, I strongly suspect that the three letter acronym to describe this file format has become generic, is permitted by an implied license, or is considered a nominative use and does not give rise to liability.
    – ohwilleke
    Jun 10, 2021 at 0:48
  • 2
    But the question assumes that the proprietary use came **first-- (which did not happen in reality) If it had, the initialism might have been protected. Jun 10, 2021 at 2:09

6 Answers 6


Being or not being open source makes fairly little difference in trademark law.

If a commercial firm (Yoyodyne, say) had used the name "Portable Network Graphics" and the abbreviation "PNG" in trade, and taken such further steps as would be needed to protect it in the relevant countries, that firm would have a protectable trademark.

Note that in some countries, a trademark must be registered to have any protection at all (much of the EU follows this rule). In others, use in commerce can offer some protection even without registration (the US follows this rule).

Had this happened (in an alternate reality) Yoyodyne could have sent a cease and desist letter when open source developers started using the mark. If the devs did not cease, Yoyodyne might have obtained an injunction, or damages for trademark infringement, or both. They could also have issued a takedown notie to the site hosting the project.

But had Yoyodyne failed to defend the mark effectively and allowed it to become generic, Yoyodyne might have lost all rights to it. Also, had Yoyodyne ceased to use it in trade for a significant period, they might have lost rights. This is a place where the different laws in different countries might lead to different results.

Note that "Portable Network Graphics" is rather descriptive, and not particularly distinctive. Descriptive marks, like "Tasty Pizza" generally get weaker protection, while more distinctive marks, such as "LuAnn's Tastee Pizza" are more strongly protected, in general. Again this depends on the country, and the specific facts.

In general the first to use, or to register a mark, gets the rights. When one entity is the first to use, but a different one is first to register, things can get confusing, and results will be different in different countries.

Note that a Cease and Desist letter is not a legal requirement. It is a threat of possible future legal action, and often an offer to avoid such action if the recipient does as the sender requests. The recipient can comply with the letter, wait for court action, or try to make some sort of compromise deal. Open source projects, as other answers suggest, often choose to comply.


For a real-world example, consider the early history of the browser now known as Firefox. The earliest releases were made under the name "Phoenix". The name was changed after trademark claims were made by Phoenix Technologies, a company that develops software but doesn't have any products even remotely related to web browsing (they build BIOS and low-level system software). The browser already had numerous public releases at this point.

The name was changed to "Firebird". This name also eventually received trademark claims, this time from an open-source database project with the same name. The developers eventually settled on the name "Firefox", which didn't have any conflicting trademarks.

These name disputes did not involve court cases, so nobody was ever officially determined to be in violation of a trademark. The trademark owner simply contacted the developers, informed them of their existing trademark, and asked them nicely to stop using a name that conflicted with it. Open-source projects typically have minimal or non-existent budgets. They can't afford to defend against a lawsuit, so a cease-and-desist request is usually honored as quickly as reasonably possible. The trademark owner might technically have the right to sue, but an open-source project doesn't generally have any assets to speak of so forcing them to pay damages is fruitless. A lawsuit is really only useful to force a stubborn project to drop their use of an infringing mark when they won't do it voluntarily. Open source software thrives on community and working with others, so historically such stubborn non-compliance is rare.

Forks, generally speaking, are a bit different when it comes to trademarks. Trademarks for names, branding, etc. are completely separate and independent from the license under which the software is distributed. Most significant open-source projects have explicit language in their licenses that explain this. You may have the right to distribute modified versions of the software, but if you do so you're required to replace the branding with your own. As a trademark holder, that means that the only forks that might also be infringing are those that aren't in compliance with the existing license, which again is not the norm in the open-source world.

Publishing software under a pseudonym doesn't really change things. It makes it a little harder for the trademark owner to figure out who to sue, but that's not a very big hurdle for a company with a legal department. Even without knowing who is behind the software, the trademark owner can submit takedown requests to whatever provider is hosting your repository and/or website.

In the Firefox case, there was a legal entity behind the project (the Mozilla Foundation, a non-profit organization). The Mozilla Foundation owned all the trademarks, branding, and source code copyrights, so the trademark disputes were pointed at them. Many open-source projects are simply a bunch of individuals collaborating without any high-level entity involved. In that case, legal disputes would likely be made against developers as individuals.

As an open-source developer, it's just as important to think about existing trademarks as if you were any other type of developer. You don't want to waste time, money, or goodwill fighting trademark infringement claims. You especially don't want to be forced into changing your project's name if you don't have to. Name changes create confusion for users and undo any name recognition that you've already earned. In the worst case, you could get sued, a judge issues an injunction that prevents you from distributing the software in any form until the case is settled, and the opposing lawyers drag out the process for years. By the time the legal case is resolved and you can get back to work, you will have lost all of your momentum, support, and volunteers. It's better for all involved if you do your due diligence and avoid stepping on anybody's existing trademarks.

You also specifically asked about users of open-source software. Users shouldn't have to worry about trademarks any more than they would with proprietary software. Trademark owners primarily go after those who create or distribute software that infringes their marks. People simply using that software haven't really damaged the trademark owner, so their ability to sue you is limited. Contrast this with patents, where many trolls absolutely will go after the end user. A patent owner can claim the end user damaged them by not paying the royalties that a proper licensee would have paid. Trademarks aren't licensed to the end user, so it's not quite the same situation.


They apply in the usual way. Infringing a trademark is not a crime, but gives the owner the right to become active to stop the usage. So, per se, Open Source developers can do whatever they want - per default, nothing will happen.

But of course they have to take the possibility into account that the trademark owner wishes to protect their ownership. For example, assume the open source project is hosted in a public github repository. The company may opt to contact the developer and send a cease&desist letter, threatening dragging them to court. They can tell them to do whatever they deem useful to their case - to remove the whole source code (close down the repo), to just remove the mention, or to rewrite the git history to make sure nothing is left. If they are especially evil, they can of course directly attach a stiff bill to pay their costs... they could even ask github to take the repository down, bypassing the developer.

Then, obviously, the dev can decide to either simply follow along with the request, or come up with a more palatable counter offer, or to deny everything and eventually risk going to court. At this point, it can become arbitrarily ugly, in the worst case a small private person fighting against the law division of a global company.

Also, you have to keep in mind that not everything is immediately a trademark infringement:

Courts consider various factors in order to determine whether a trademark was infringed.

[...] Whether the defendant's use of the mark is "in commerce." Whether that use is connected to the sale, offer, distribution, or advertising of a product. Whether the defendant's use of the trademark is likely to confuse consumers. This last factor, consumer confusion, is the main topic of debate in most cases.

I feel very hard pressed to remember any Open Source projects where these kinds of points had any relevance.

In most cases, open source software is not sold - at least not in competition to the trademarked version - or particularly advertised; and it should be easy for the developer to make clear that their software is not from the trademark owner, hence unlikely to "confuse consumers".

Software which is in direct competition with a commercial product has no real need to mention any trademarks (unless it wants to pretend that it is related to the competition - but I guess that's a case you're not asking about). Other software often thrives in the "biome" of a large commercial software; for example there is a huge ecology of little tools, software libraries and so on to make commercial RDBMS like Oracle more accessible. Those are, in 99% of the cases, filling holes the trademark holder has left in their offering, and while they may use the trademarked name, arguing any of the aforementioned critical points in court may be very hard indeed for the owner.

At the end of the day, these issues will go to civil courts, not criminal, so it is always possible that the holder would just try to make life for the developer as uncomfortable as possible (if not outright winning in court) and thus trying to protect their ownership. That, to me, seems to be the more threatening thing than the actual court ruling (and - compared to patents - really unlikely/rare).

  • 2
    What complicates this is that many open source projects do not have an owner at all; instead, each individual contribution is owned by its contributor, or the entity they work for at the time they made the contribution, so when the owner of some right (trademark, copyright, patent, personal info (GDPR), whatever) thinks the code infringes on that right, who is liable? Whoever first introduced that code? Whoever ever worked with that code? Whoever uses the code? Whoever publishes it? This is the aspect of the question that interests me, this is the real question, and I have no good answer. Jun 10, 2021 at 11:26
  • @reinierpost, I don't think copyright comes into that. You would go after the "artifact" (the code or the documentation using the trademark...). If you cannot find a good primary developer/owner for the software, then nothing keeps you from going to the entity providing the physical access to the software (in my example - github, or other repositories). You can easily ask them to remove all forks of the software... finding them is part of your job as trademark owner; and there certainly are law companies specialized in that. If you are successful - who knows...
    – AnoE
    Jun 10, 2021 at 11:35
  • copyright is just another example of a right you may own and want to enforce. yes you can go after the publisher but there may not be just one, there may literally be thousands. getting a repository removed from GitHub in no way guarantees that the code or products built from it are no longer out there Jun 10, 2021 at 11:59
  • 1
    Sure! Companies have been known to employ law subsidiaries with the mandate of going through all occurrences out there and shutting them down individually (if not for trademarks, then certainly for patents, software piracy and so on). It's a real, small, industry out there... It is not necessary to "guarantee" to catch all of them.
    – AnoE
    Jun 10, 2021 at 12:04
  • 2
    The answer said "In most cases, open source software is not 'in commerce', nothing is sold or particularly advertised". But "in commerce'" can mean something that is offered or distributed to the public free of charge, particularly if it competes with products being sold. A disclaimer of relationship may or may not be effective to avoid an infringement claim. Jun 10, 2021 at 12:43

When Debian was adding their own patches to Mozilla projects when compiling them for use with the Debian OS, Mozilla asked them to either stop adding the patches or stop using the Mozilla-owned trademarks. So in 2006 Firefox was re-branded to Iceweasel when the open source project was compiled and bundled with the Debian OS.

Similarly, Thunderbird became Icedove and SeaMonkey became Iceape.

After a decade of this, Mozilla decided that the re-branding wasn't needed anymore, and Debian is back to the regular names now.

See more details: https://en.wikipedia.org/wiki/Mozilla_software_rebranded_by_Debian


This question pretty much goes to the heart of trade mark law, and, as such, is pretty complicated and, to say the least, books have been written about it. The crux of your question is “what is a trade mark, and why does it exist?”. You may not feel you are asking that question, but that’s what’s required of a proper answer.

A key part of trade mark law is that applying for a trade mark involves fitting your claim with the Nice Classification categories. Effectively that means that if you register a trade mark for cheese someone else can use the same trade mark for an aircraft model or mobile phone (with some exceptions).

In addition, the trade mark you register will have to some distinctiveness to it, it cannot be purely descriptive. Hence you often see trade marks like Norfpol or Kulor rather than North Pole or Colour.

Referring back to your question, it is arguable at this point that Portable Network Graphics is merely descriptive, so possibly not eligible for trade mark protection ab initio.

“Ah,” you say, “but what about all those trade marks that are not really distinct, and are descriptive, or perhaps just place names: Vauxhall, Mont Blanc, Columbia, Apple, The North FaceTwitter, Facebook, Yahoo

In these cases the applicant has made the claim that they have acquired distinctiveness, and that their use of the mark is distinguishable from the common use of the term.

So, possibly, Portable Network Graphics could acquire distinctiveness, and be free of possible challenges to its trade mark registration. On this point, there are armies of lawyers watching trade mark applications, ready to challenge those they deem to be too close to those of the clients they represent.

Finally, the real test is whether there would be confusion in the minds of the public, between different uses of the same mark by different companies. I am sure that there are plenty of examples of this, but the household cleaner Cif was formerly known as Jif in the UK. It so happens that at the same time there was (is) a brand of concentrated lemon juice also called Jif. I don’t think there was much confusion between the two sets of customers.

@WhiteHotLoveTiger makes the point about open source software and the Mozilla/Firefox marks. While this point may have been worked-out with Debian, there are other Mozilla-based browsers that do not and cannot use the marks.

So, I hope that you can see that the question you have asked is contingent on many factors, and the complex web of this area of law has been worked-out over hundreds of years.

  • I hope the logos of the household cleaner and the juice were reasonably distinct, otherwise there could be really dangerous confusions. Jun 12, 2021 at 11:25
  • If they were in similar packaging in the same area of a supermarket, there could be a big problem. Likewise, just asking someone to "get me some Jif" could also be problematic. :) In this case the packaging of the lemon juice is lemon-shaped/sized, so not much risk of confusion. If I recall correctly the packaging of the juice did turn into a trade mark dispute – someone else wanted to use a similar device. Just one more note here: trade marks are not just graphic/logos, they can be words. There have been attempts to register colours and sounds – most, if not all, of which have been rejected.
    – typonaut
    Jun 12, 2021 at 19:01

This wiki was create by the asker to have a simplified version of information provided by other answers, to inform a general audience of developers. Any edits are welcome with this goal in mind.

  • Trademarks are still trademarks. Trademark owners are responsible for protecting their own marks by pursuing legal action.

  • In order to find you liable in court, the trademark owner would need to demonstrate that you had used their mark in some infringing way (namely, in a way likely to create customer confusion associated with their brand).

    • Whether or not one's use of the mark is infringing could depend on multiple factors including whether it was used in commerce and whether the mark potentially made an impression on consumers (wherein the confusion would arise).
  • In reality, of course, the trademark owner doesn't necessarily need a plausible case in order to pursue legal action, only to get a favourable ruling.

  • Since the legal budget of open-source projects is likely near 0, the trademark owner may see this as an easy opportunity to intimidate the developers into re-branding or removing the project. At the same time, they may see that the amount of damages they would be able to recover from you is so small, that it's not worth the cost to pursue.

In the case that the trademark owner takes issue with the existence or distribution of the code itself, rather than a particular use of it in the market, the most likely outcomes are the following:

  • The trademark owner will contact the developers, asking to cease distribution of the software, or change the name. A range of outcomes could result from this, as discussed below.

  • The trademark owner will make a formal take-down notice to the site(s) hosting the software, such as GitHub. The host will most likely comply with this request immediately after confirming only the most basic merits (and we can't really expect more from them). In most cases, you can submit a counter notice to re-instate your project, which maybe granted if the first party doesn't continue to pursue. You can see on GitHub that over 250 DMCA take-down notices containing the word "trademark" have been received.

In general, it's best for all parties to avoid potential naming conflicts. Even if there is no legal basis for a specific punitive action, the naming conflict can still cause inconvenience. A legal battle will likely cause unneeded stress and resource drain on the project. If people associate your project with lawsuits, they probably will be less keen to use it. Even if you comply with the other party's request to change the name, actually changing the name could be a very challenging task and may kill the momentum of your project after confusion and shocked productivity.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .