49

I coded the methodology based on the paper "Free Water Elimination and Mapping from Diffusion MRI" and I am curious if I am legally bound to take permission from the authors or the journal in making the code open source on Github?

  • 7
    It looks like you've accidentally created two accounts; follow the instructions here to merge them so that you can edit and comment on your own question. – ff524 Jul 11 '16 at 15:39
  • 8
    If you do open-source it, you'll do a great service for research. Thank you! – Ric Jul 11 '16 at 15:52
  • 3
    You need to start by reading en.wikipedia.org/wiki/Copyright and en.wikipedia.org/wiki/Patent. That will help you ask a more informed question. Questions are better when you do research before asking, so you can ask an well-informed and well-scoped question. As it stands your question is too broad -- you're basically asking for an overview of intellectual property law, which seems like too much to ask for on this site. – D.W. Jul 11 '16 at 22:15
  • Also, if you do have a more specific question, you might find a different set of useful comments on the Open Source SE site. – MAP Jul 11 '16 at 22:23
  • 7
    You should still contact the authors of the original paper and let them know. They will probably be happy to know somebody has brought their methods from theory to practice! (And they will probably point interested researchers to your GitHub code, too) – elisa Jul 12 '16 at 9:02
75

You are not required to get permission to implement an idea you have read about in a paper, or to make it open source.

The paper is protected by copyright, but this only protects the text and images in the paper (the expression of the idea), not the idea itself. The copyright on the paper does not prevent you from creating your own realization of the idea in code and distributing it freely. You should also acknowledge the source of the idea by citing the paper in which you read about it, but this is an ethical requirement, not a legal one.

The exception to this is if the methodology in the paper is patented; in this case additional restrictions on what you can do will apply.

  • 39
    Software and abstract method patents are evil, so we hope scholars don't do that, right? *ducks* – cat Jul 11 '16 at 16:45
  • 29
    The last paragraph depends a lot on the jurisdiction. In EU no software patents exist. – Bakuriu Jul 11 '16 at 17:05
  • 8
    So your answer is actually "You might be required to get permission; we can't tell you, as that depends on whether the scheme is patented or not". That's pretty different from the summary in the first line of the answer. The exception at the bottom of your answer is an awfully big one... – D.W. Jul 11 '16 at 22:16
  • 5
    @Bakuriu: A good lawyer can write a patent that is not a "software patent" and protect something that looks exactly like an algorithm this way. Even for algorithms, you really need to determine whether they're patented or not before you release the code. – Peter Shor Jul 12 '16 at 0:54
  • 4
    It might be nice to point out that it would be polite to ask permission even if it is not required. Or simply thank them for the paper and state that you have attempted to implement their idea and would welcome feedback. Acknowledgement somewhere in the project or source code would also be polite. None of that is likely to be a legal requirement unless the paper has a license on it that mandates it. It is what many may feel would be good academic practise though. – TafT Jul 12 '16 at 10:44
5

IANAL and thus I won't give legal advice. Whether you have to ask them or not is irrelevant. Asking them is a courtesy which the authors will most certainly appreciate and might even cause them to endorse your project, maybe even as the official implementation which they might also contribute to.

In the unlikely event that they strictly deny you permission (maybe they already have their own implementation that they want to distribute commercially and thus they wouldn't want a free competitor) then is the correct time to take legal advice. Assuming you're working in an academic institution they most likely have a legal department which you could and should check on. Of course you can also do so before contacting the authors, maybe asking about the proper approach to not make you accidentally liable.

Just don't trust the internet for legal advice.

4

Might be best to check whether the information in the paper is covered by more than just copyright. In the US, for example, even mere algorithms can be patented, so implementing the algorithm in code, even though the code itself is all your work and copyrighted to you, can still be a patent violation.

  • 2
    I don't think the code itself violates the patent. I think actually running the code may be a patent violation. – Joost Jul 12 '16 at 8:51
  • 2
    @Joost: writing and running code for personal use that uses information from a patent should be fine. Distributing it (even as open source) could be a patent violation. Publishing images produced using it would be a grey area, but probably a violation. Using the code for non-personal use (ie, in just about any organization which might conceivably have access to MRI imaging) almost certainly would be a violation. IF there's a patent in the first place. Definitely something to ask a patent lawyer about, if there is. – Dewi Morgan Jul 12 '16 at 13:50
  • 4
    @DewiMorgan Other open source projects, such as the LAME mp3 project , distribute source code that if compiled will preform a patented process, but their position is that the source code itself cannot violate the patent as it is merely a description of the process and thus they distributed it freely under an open source license. – Dev Jul 12 '16 at 23:02
  • @dev If your lawyer feels that's defensible in your jurisdiction, that might be viable. Legal advice MUST be the first step. Chances of LAME devs' arguments standing up in any technically literate court are slim to none, especially since they give links to compiled binaries: there's no "but we're using download.com!" defense in court. Also, if their "but it's only source code" argument were used on a port to an interpreted language, it'd be nonsense: and C interpreters do exist (CInt, Ch, etc). They only aren't sued as it's more profitable to hit up their users for license fees. – Dewi Morgan Jul 13 '16 at 18:10
  • @DewiMorgan Ninth Circuit has already held that source code is expression. It's protected by copyrights not by patents. What it does may be patented but you are not performing a patented process by distributing a description of it which is all source code is. – Dev Jul 13 '16 at 21:36
-3

IMHO you must cite their paper in your Github repository explicitly, but you are not required to get a permission.

  • 20
    This is an ethical and professional/academic obligation, but given the very explicit wording of the question, we should be clear that it is not a legal obligation. – Phil Miller Jul 11 '16 at 17:02
  • 7
    It is neither an ethical, professional, academic, nor legal obligation to use Github for your open source project. – emory Jul 11 '16 at 19:23
  • 2
    @emory: The original question said it was going to be on github. I don't think the answer was saying the obligation was to use github, but that since you are using github, the obligation is to cite the source in your github repository. – MAP Jul 11 '16 at 22:19
  • Although the question asks for the legal obligations and not for the academical ethic standards, I suspect the second has a much stronger effect to the everyday practice. In academical circles, criminality is nearly unheard, violating the second is not. Thus, on my opinion, question doesn't focus the real problem. – peterh says reinstate Monica Jul 12 '16 at 2:12
  • Actually, if the legal situation is hazy, citing the work on GitHub may be asking for trouble (regardless of ethical considerations). – einpoklum - reinstate Monica Jul 12 '16 at 11:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.