I should be asleep, but I accidentally came across this article calling Trademarks a hidden menace to Open Souce. At it’s heart I agree that Trademark law is not properly understood by the community at large and could be a major problem for some projects. Beyond that it’s is just a mess.

The start of this polite rant was that he is the author of a commercial Unbuntu book, and he made a website to support the book. He was clear in the use of the trademark ownership, but did not get permission first from Unbuntu. During conversations with Connonical it was mentioned that he might have taken liberties, but that was just in passing. Instead of asking for permission (which we would most likely have been given) he removed the logo and the trademark in all places except where nominal use could be applied (i.e. the book title) and writes a blog post calling for the abolishment of the use of Trademarks in Open Source.  To me this is Trademark law working perfectly well for Open Source. If you are doing something commercial, the owner of the mark wants the ability to aprove it to make sure its not malicious. They are taking as light a hand a possible in it’s use in allowing non-commercial use freely without the need for seeking aproval (NOTE: you have to use it.. no really, read on)

There are a number of mental failings in that posting. First and foremost is that it forgets the lesson that the GPL and Open Source are reliant on the copyright system (and I would argue that even if there was not copyright that I would want it if only FOR the GPL which enforces the principals of giving back.) The argument is not to have any trademarks at all (as no other option is provided anywhere.) He calls out trademarks as being the antithesis of Open Source. You could literally replace ‘Trademark’ with ‘Copyright’ in his end section (with one minor exception.) But we need copyright law for the GPL to work; its better than the Public Domain. Just as with Copyright, it is the application which should be addressed, not the tool it’s self. Any ’solution’ should use the trademark system to support Open Source.

What tends to happen is that open source companies have to walk a tightrope, and slightly strange rules on trademark get put in place. For example, Ubuntu is cool with community remixes using the trademark, but if you intend to make money from Ubuntu and want to include the word in your business title, you’re going to need permission. It’s not quite clear here how the former won’t dilute the Ubuntu brand, while the latter possibly will. The “protecting brand identity” argument falls apart almost immediately upon examination.

Um… then you do not have much experience in the real world. Would a company which called its self ‘The Unbuntu Corporation” and sell a product ‘Real Ubuntu’ based on a very old release or containing problems on purpose dilute the name? What about the ‘Python Corporation’? What about someone selling ‘Django 2.0′ (which turns out to be a fork based on 0.95). Would these things dilute those projects? I think they do. Allowing free use for non-commercialuse may have a risk of dilution, but as this is an Open Source project and crucial to the founding and flourishing of a community, more often than diluting, it helps build the brand. Those are the uses which you want to promote as being part of an Open Source community. These are real issues and real problems which we are facing today.

Here is the crux. If you choose not to use Trademark law to your benefit, others can use it as a weapon against you. If you do not actively protect your trademarks, you loose them, and you can be attacked. I actually am a little happy that there have been some recent issues with the use of the Python logo and trademark as it proved that they are important to have and protect.

The nature of Linux and open source in general is to encourage forking and splinter projects. That’s the basic freedom provided by the GNU Public License, and similar licenses. Some of the forks or splinter projects will be poor quality. Some will fail. But that’s just the way things work with Linux.

Ah, and here’s the rub. If a project forks into 2, or say 4 different projects, do all those projects have the same name? What if all the different linux distributions were just called ‘Linux’ without the qualifiers ‘RedHat’, ‘Fedora’, ‘Unbuntu’, etc. Branding is important. If you invent a better version of FireFox based off of the source code but with different options and behavior and code, then do you get to call it FireFox? Would you even want to? Would you want it to be confused with this ‘failed’ version? Wouldn’t you want to rebrand it to call it out as being better? Maybe IceWeasel? Yea I know I am going to get flamed. I agree that debacle was stupid. But I can understand how the problem came about and I do not see malice or ill intent as the author does:

Mozilla is even worse. If I create a new Linux distro, and include my own compiled Firefox binary, it’s unlikely I would be able to call the browser “Firefox”, or use the familiar fox logo, without getting permission from Mozilla. This could put me at a competitive disadvantage compared to other versions of Linux because my users would be using what appears to be unfamiliar software. It’s worth mentioning that Mozilla’s trademark rules also indicate they’re not terribly happy about the unofficial redistribution of their binaries, either, and would prefer it if they were the exclusive source.

But its not the same software, now is it? (well maybe not in this exact instance, but that is what the Trademark protection is supposed to be used for. To prevent a fork from being called the same ‘proper name’ as it is a fork.) Once again it is the use of Trademark law, just like copyright, that matters.

Is this how open source is supposed to work? Redistricted redistribution? Tight control on who can compile software and still be able to call it by its proper name?

Um.. yes and no. There is no restriction on the distrobution of the source code. Trademark has nothing to do with source code, it is just about the use of a word or words as an identifier. There is nothing stoping people from compiling software can calling it by its given name. It’s not about the ‘who’ but about the ‘what’; or at least it should be that way. If I compile firefox exactly like the Mozilla Foundataion does, then I can call it FireFox. If I change what could is compiled and packaged, I can not call that package ‘FireFox’. I do not agree with this decision myself, but I do understand the slippery slope they are trying to protect against.

So lets recap:

  1. Trademark in and of it’s self is not the problem (the same way copyright is not), but how it’s applied.
  2. If you do not use Trademark Law, it can and will be used against you (and can be usd to make you change your existing project from say ‘Phoenix’ to say ‘Firefox’ and pay fees for violation.)
  3. Trademark law does not impact your ability to distrobute, change, fork or compile code.
  4. Trademark law does impact what you can call something. It gives you the ability to say how a name can be used.
  5. The improper use of Trademark Law could be a problem for Open Source, but the majority of the issues (and the one which started his rant) are actually good uses of Trademark Law for Open Source.
  6. Preventing corporate and malicious use while promoting the non-corporate communities, and approved corporate use is good for Open Source.
  7. Why not ASK if you can use the trademark on your website for your book promoting the project?

There is a part of me which would love to take about 20K and start to go all evil with respect to trademarks, taking Open Source names and preventing projects from calling themselves things. Forking projects and trademarking them for my own purpose and calling them the ‘REAL’ version. It is not about not Trademarking, but about comming up with good guidelines for its proper use in Open Source. It is a tight line one must walk, but walk it we must.