I am currently writing a conference paper (an IEEE conf, if that matters) in which I include a piece of code. Pertaining to the license of the code, should I, at the beginning of the code, write something like

-- Author: Ton Ami, 2021.
-- License: MIT License. ------> this
--  Below is code to help you make a better world

Or should I just put the code without any license terms?

In any case, what would the license of the code be?

Related questions:

License of code accompanying a published article

Is there a default license for code libraries that accompany scientific research papers

Include code snippet which is licensed under the MIT License in a paper

  • 1. How long is this code? Two lines? One page? More? 2. How is it 'included'? Directly in the main text? As a separate file? Commented Mar 3, 2021 at 8:17
  • 1
    Is the code yours? (I assumed so in my answer. That might not have been your intent.)
    – user2768
    Commented Mar 3, 2021 at 8:39
  • 1
    @Louic I disagree; maybe it would not change your answer. Commented Mar 3, 2021 at 9:55
  • 1
    @FedericoPoloni There is no such thing as "my answer". If answers are opinion based the question should be closed, but that is not the case here.
    – Louic
    Commented Mar 3, 2021 at 10:01
  • 2
    @Louic I disagree on this, too. However, feel free to add an answer if you think there is sufficient detail to give one. Commented Mar 3, 2021 at 10:04

4 Answers 4


If you publish something and don't give an explicit license (US, anyway), the default is "all rights reserved". You can provide the license elsewhere if you like, as long as you still hold copyright. But if you transfer copyright to someone else, and it doesn't already have an explicit license, then they own copyright and you can no longer provide a license.

If you do provide a license before copyright transfer then, for most such licenses, the license is "sticky" and can't be withdrawn, even by the new copyright holder. If you do that without informing the new copyright owner then you can be in some trouble, since you can no longer transfer "all rights", but only the limited rights you still hold.

Some publishers, including conferences, will be fine with this. Others not and you could wind up having a paper rejected/withdrawn. I suspect IEEE might be lenient here, but they need to be informed of any prior license to anything they publish. Putting it in the paper itself makes it obvious, of course.

However, it isn't where you provide the license terms (in the paper or elsewhere), but when you detail the license terms, assuming that you can, at all.


I would suggest that you deposit your code (with whatever license you want) in zenodo and include a link to it from your paper, in addition to including it in your text.

It doesn't matter what the conference copyright transfer is if your code is released under MIT. I guess they could refuse to take your paper as a result but that's never happened to the best of my knowledge. As there is no provision for revoking the MIT licence, even if you give them the copyright, you'll be able to do anything you want with the code because it is licensed to you under the MIT license (including continuing to distribute it), as will anyone else who obtains it from zenodo.


Technically, you should include the full MIT license at the head of any major source files in your code. (here's a link to the full license text). Since you are using the MIT license, which is incredibly permissive, it functionally doesn't make much of a difference if you include it or not in all your source files.

This is assuming you are posting a link to the actual code (e.g. to the github page or personal website etc.). If you are posting the code in your paper, definitely do not include the license there, as no one ever does that.


Ask your university's lawyers. They'll need to check whether the publisher's copyright agreement attempts to overreach, perhaps resulting in the inadvertent transfer of rights to the publisher.

Alternatively, include code extracts under fair use, attributing yourself as the programmer (rather than the author), removing any claim that the code is a contribution, and mentioning somewhere that the code is under MIT License. (Strictly speaking, university lawyers should still check whether you've complied with fair use and not violated the MIT license.)

I assume code was written by the OP.

  • 2
    "publisher's copyright agreement attempts to overreach, perhaps invalidating any claim that code is licensed." In my experience (with journals not specializing in software) this is always the case and is plainly stated. Commented Mar 3, 2021 at 8:40
  • @AnonymousPhysicist Publishers clearly cannot reasonably claim to control copyright of code included under fair use, so why do publishers try?
    – user2768
    Commented Mar 5, 2021 at 6:44
  • @AnonymousPhysicist Also, assuming you're right (overreach is always the case, licensing claims for code are invalid), then code can never be published (without assigning copyright to publishers), unless code extracts are include under fair use (i.e., code isn't a contribution of a manuscript, code belongs to an independent project, governed by independent copyright/license.)
    – user2768
    Commented Mar 5, 2021 at 6:59
  • 2
    This has nothing to do with fair use. Closed access journals generally demand the copyright to the document. If you do not give them the copyright, they will not publish it. Commented Mar 5, 2021 at 7:27
  • 2
    I think that a number of commenters here don't understand "fair use". It isn't the same as "public domain" and is, instead, a limited exception permitting some copying of copyrighted material; usually short segments that don't affect the (monetary) value of the whole. The author doesn't "release under fair use".
    – Buffy
    Commented Mar 5, 2021 at 14:25

You must log in to answer this question.

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