Can programming language names be trademarks?

Register or Login to like
Register or Login to like
Trademark symbols

Can–or should–a programming language name be a trademark? The Trademark Trial and Appeal Board, the administrative board within the U.S. Patent and Trademark Office that adjudicates whether trademarks can be registered, recently decided that the word “Lua” was not a generic name for a programming language, but rather that the term referred to a particular proprietary language. Is it right though, that someone is allowed to exercise proprietary rights over a programming language name?

A word is only a trademark if the consumer recognizes it as one. We all know the litany of former trademarks that aren't trademarks anymore because people used the words to identify a type of goods, not a particular manufacturer's brand: aspirin (in the U.S.), shredded wheat, cellophane, escalator, and thermos are a few examples.

So do people think of the names of programming languages as trademarks? For some programming languages, there is clearly no single entity that claims ownership. Fortran, Ada, C, C++, Extended Pascal, and Basic are all defined by ISO standards and no one owns the names; they're just generic. Based on past experience, as users we may not be predisposed to think of a language name as proprietary.

But there are also a fair number of trademark registrations for computer language names, like for Javascript, Java, Python and Perl. But these aren't registrations for language names per se; rather, the registrations claim computer programs related in some way to a language, like this example for Python: “computer programs and downloadable computer programs that implement an object-oriented computer programming language.” The absence of registrations for programming languages per se is perhaps on the theory that a language is neither a good or a service that can be used in trade, but rather an abstraction. Instead, a program that assists in programming in a given language, like an IDE, or a book that describes the language, are the tangible goods that can be distributed and therefore have a trademark.

There are more indications that a language name is simply not a trademark. The USPTO says that in a trademark registration, one may only use generic words to describe the goods and services – it is “inappropriate” to use a trademark in the description and the examining attorney should instruct the applicant to change the trademark to a generic word. Nevertheless, a number of registrations contain the trademarks “Java” or “Perl” in the description of goods and services, e.g., “M Squared Technologies” is a long-standing registration for “computer software development tools that assist the software developer to create and understand software programs written in the computer programming languages of C, C++, Java, C#, Fortran, Pascal, Assembler, XML, HTML, and variants of Basic.” This uses the “Java” registered trademark right alongside the generic names for other languages. “J2U” is registered for “conversion of software into Practical Extraction and Report Language (PERL), korn shell (KSH) or Bourne-Again Shell (BASH) command language interpreter,” using the Perl registered trademark. This certainly suggest that at least the trademark examining attorneys aren't recognizing these words are trademarks. Another example: when Microsoft filed its application for Visual C#, it disclaimed the “exclusive right to use C# apart from the mark as shown.” Microsoft isn't claiming any exclusive rights to the name C#.

Does it really matter though? And isn't it fair that the entity that controls the standard for the language be allowed to say what is authentic $LANGUAGE and what isn't?

It does matter. No one should be able to use trademark law to control truthful speech about a product. The Examining Attorney in the Lua application raised this point, but it wasn't a good enough sell for the TTAB. Indeed there is a legal theory – nominative fair use – that says one can use others' trademarks to refer truthfully to the others' products, but this is a judge-made doctrine, not statutory, of varying parameters in different courts, and not even accepted by all of them. Given the unsettled law, one can see that an aggressive trademark enforcer could exploit the situation to try to suppress use of the programming language itself.

Pam is a Board Member of the Open Source Initiative and the principal of Chestek Legal in Raleigh, North Carolina. She works with creative communities, giving practical legal advice on branding, marketing, and protecting and sharing content.


Perhaps I show my ignorance, by it has been my impression that Trademarks are less dangerous than Patents. No matter what word is trademarked, you can always use a creative alternative. It's not a Yacc™ it's a Bison. If I am missing something, I'll be pleased to be set straight.

That said, there *must* be *some* public domain word recognized for referring to any abstraction. Otherwise, you can't talk about it. It would be common courtesy (and common sense) for any business to set aside generic terms while trademarking their concrete realizations. It sounds like that is what Microsoft did by leaving "C#" available while trademarking "Visual C#™".

The real danger (to users of alternative realizations like Mono[™?]) is if they succeed in patenting the language.

Microsoft already patented .NET and C#. Mono is written by De Icaza specifically to infringe Microsoft patents so that Microsoft can claim they "love open source" because they are not suing us (yet). When they infect GNOME with Mono so that infection can't be easily removed, they will sue. Probably using some of their patent trolls like Paul Allen.

Good thing is that Red Hat doesn't ship Mono in Fedora or RHEL. Bad thing is that Canonical does.

If a language has a trademark, even referential use could still be a problem. What if I said "I wrote this in MyLanguage." MyLanguage is the trademark owned by MyCompany, and I get a letter from MyCompany saying "You're not using authentic MyLanguage because it doesn't meet our standards, so you can't say that." But how else could I convey that I used the language everyone else thinks is MyLanguage?

I'm not sure whether a generic identifier solves the problem because it just seems duplicative. But it is interesting, I think, that we commonly refer to a "generic [name brand]" for pharmaceuticals, like "generic Advil," and no one seems to worry too much about it. I'd love to hear from someone familiar with pharma trademarks about the generic industries' use of brand names.

From 1979 to 1987 <a href="">when the DoD let the registration lapse</a>.

Unidentified - awesome, thanks!


I was there in the early eighties, and ADA was one of those 'Official" DoD languages, like Modula2.

What's the beef? Your argument seems to hinge on a vague accusation that a trademark holder can stop someone from using the trademark. Maybe I'm misunderstanding but the common case is that if you want to legitimately refer to a trademark'd name, simply append the trademark symbol and IIRC also include a statement somewhere that such-and-such is a trademark of so-and-so. What's the problem?

See the comment from John Roach. You are right that there should be no question that these kinds of uses are permissible. But the nominative fair use doctrine, which covers these situations, is not accepted by all courts and varies in scope, so there is a lot of gray area. This means, like in John's case, that third parties are overly cautious, and that the trademark holder itself can exploit the ambiguity with suit or threat of suit. Trademark owners are pretty well known for trying to suppress speech that they don't like through a trademark infringement claim.

Calvin Moorers (

"He founded the Rockford Research Institute in 1961, where he developed the TRAC programming language, and attempted to control its distribution and development using trademark law and a unique invocation of copyright.[2] (At the time patent law would not allow him to control what he saw as his intellectual property and profit from it.) The trademark strategy was later used by Ada."

and TRAC (


Me and my professor wrote a paper for IEEE Consumer Electronics. We were expecting an acceptance for our paper when we learned that our paper was denied. It turns out that writing the language name of your program is not allowed. (FYI I was using Python) This shocked us. However we changed the "Python" to "a object oriented programming language". Which I must say I felt nauseated. Python isn't just some "object oriented programming language".
We couldn't use the following names for our paper;
- Linux ( I mean Linux!!! Come on... It's free! And you can create your own distribution with no problems! RedHat even sells their own Linux! )
- Apache ( They may be right with this one.)
- Python (Read above)

This is where creative naming comes in. If you aren't allowed to use the trademarked word, then use another, and define it. So at the beginning of the paper, briefly describe the language "Monty", and then use that to refer to the language. You can even have references to "Monty like language implementations" and reference C Python™ Jython™ and such in the Bibliography. (Pardon me if Monty is already claimed.)

It is a ridiculous exercise, of course. We need something like Creative Commons for trademarks. When a company or foundation has a product like Python, there should be an explicit license, chosen from a small pre-written list, to authorize generic public discourse.

Another problem is that even if there *was* such an explicit license, the people rejecting your paper might refuse to recognize it. In which case, they are the "ijits".

I'm sure there are hundreds of examples of published IEEE journal papers making exactly the kinds of references to specific programming languages and software that you mention.

Yes there are lots of examples. We even pointed them out. However they did not budge. I wonder if I can copy and paste the e-mail I got from "WAYNE C. LUPLOW Editor-in-Chief IEEE Transactions on Consumer Electronics" and not take fire from them. However I will paste a quote;

"Per our Instructions, the use of commercial names such as Linux, Apache and Python is not permitted."

I wish it were not true. But this is what we got...

Sounds like advice from an overly conservative lawyer. Maybe that's the bigger problem, the impairment of ordinary discourse because the law isn't clear enough, not threats from the trademark owners themselves. .

Yes, I've noticed that trademark owners usually have no object to their name used in general discourse. They send a polite but firm letter if their name is (or seems to be) used in advertising (like the letter from Hormel™ to Think Geek™ for their canned unicorn meat ad). They only really get upset when their name is used in parody (like Barney™ parodies). Parody is protected, but it takes expensive lawyers to remind them of that. In parody cases, corporate trademark owners need to find a way to grow a corporate sense of humor. :-)

Yes, I've noticed that trademark owners usually have no object to their name used in general discourse. They send a polite but firm letter if their name is (or seems to be) used in advertising (like the letter from Hormel™ to Think Geek™ for their canned unicorn meat ad). They only really get upset when their name is used in parody (like Barney™ parodies). Parody is protected, but it takes expensive lawyers to remind them of that. In parody cases, corporate trademark owners need to find a way to grow a corporate sense of humor. :-)

That seems likely, especially given those specific examples (Linux, Apache, Python). I am sure the owners of those trademarks would not object to the kind of ordinary references that are really unavoidable in technical papers discussing the use of such software or programming languages.

You wrote, "..a type of goods, not a particular manufacturer's brand", and wrt Thermos, the word eventually entered (or was allowed to enter) the generic term category.

"Whoa dude! He must have written that program in Lua Brand Lua!" LOL!

"I wrote this in ANSI C" - hm... okay, er... that's not going to make a great point now is it? Unless the American National Standards Institute decides to Trademark their spec on the C language....

But, "You wrote that crappy app in Borland C" or "All apps written in Microsoft(r) VisualBasic(r) are total junk!" - yeah, these might illustrate your argument here Pam :)

Another case in point of how avoiding loss of trade/servicemarks can sound ridiculous: "I am stuck on Band-Aid [Brand] coz Band-Aid's stuck on me!".

Back in the seventies or sixties, I don't think the word, "Brand" would have been needed, but since there are now Curad and a host of other makers of "Band-Aids" (sic, okay so they're, 'bandages'), Calling your product, "A Band-Aid", instead of, "Band-Aid brand bandages" ensures that not only will the consumer continue to call all bandages dispensed from little cans, "Band-Aids", but also, that you ain't gonna lose your trademark.

Also, "Kleenex Brand tissues", instead of "Kleenex's".

Hey folks! We're still going to call Kleenex tissues Kleenex's, but we're never going to ask for a paper towel by saying, "Hey can you tear me off a Brawny?"

There is one major issue w/the trademark lobby, however. A few years back a change in IP law took place, and if you don't "zealously defend" your trademark, you may lose it. Hm...

Okay that brings me to the title of this talk-back: the "Flattening".

When you talk about domain names, let's say, for example, that you are Volkswagon... No, let's say you are the ISP, Virtual Works.

According to the Internet RFCs, you should register VW.NET, since you are primarily a network, right? Right!

Yes, that is what they did.

So, if you were Volkswagen, you should attempt, on an FCFS basis, to register VW.COM - makes sense, or maybe

So one day, the Atty's for Volkswagen say to the upper brass, "If we don't zealously defend our trademark, we can lose it.".

So Volkswagon (or is it wagen) approaches the ISP, Virtual Works, about purchasing their VW.NET domain. Not Interested was the response from my good friends at Virtual Works.

Then, Volkswagen comes back and says something to the effect of, "We'll give you a zillion bucks to transfer that domain to us, and with that money you can re-brand your company and have plenty of money to spare!"

"A Zillion Dollars? Heck yeah! We'll do it" (You would too unless you know where this is going).

Volkswagen rescinds the offer, with a much more modest number of, oh, I don't know let's say, a hundred bucks for the domain.

Virtual Works says, "Huh? are you nuts? NO!!!"

Volkswagen picks up the phone, calls the newly created ICANN and invokes the UDRP with the support of WIPO and the trademark lobby, now that those two special interest groups have successfully lobbied for an incorrect definition of just what "cybersquatting" really is...

Voila - Virtual Works, and all their customers, have disappeared from the Internet forever.

Well my friends out there in intellectual property land - the DNS protocol is desgined to be a taxonomical, distributed tree of insignificant characters (to a machine) that are interpreted mnemonically by humans, and translitterate those hierarchies of strings into an address system of numbers.

And guess what, you've just effectively FLATTENED the DNS name space - with that sort of legal opinion running rampant in th IP world, mixing it with domain names means that either:

1.) you go bankrupt because you have to usurp and/or register your trademarked string in each of the hundreds of extant gTLDs and ccTLDs

2.) You realize that this is futile and juvenile, and opt instead to just eliminate all ISO-3166-2 ccTLDs for the various sovereign nations of the world along with .NET/ORG/MIL/GOV/US/INFO, etc..., and only have .COM

Yup. so... is trademark law and the intellectual property special interest group going to flatten the English Language as well?

=== What actually happened to Virtual Works: ===

When Volkswagen invoked UDRP under the fledgling ICANN's procudures, the domain was stolen from Virtual Works through a process we call, "Reverse Domain Hi-Jacking" - meaning, that ICANN received a complaint from Volkswagen alleging their Intellectual property was being held as a hostage - for a ransom (of a zillion dollars, mind you - remember the initial offer? Volkswagen didn't have to disclose that it was THEY, who initiated the offer, and by an act of omission, implied that Virtual Works was ransoming Volkswagen's trademark).

All without Due process.

Trademark law allows for very severe penalties for infringement, yet ICANN and the Intellectual Property interests have effectively done away with due process in favor of an arbitrary system of favortism toward those who fund ICANN.

=== What actually happened to Virtual Works: ===

Great article Pam!

Kindest regards,

Creative Commons LicenseThis work is licensed under a Creative Commons Attribution-Share Alike 3.0 Unported License.