25

As a postdoc I joined an existing project of a PhD student. I made significant contributions to his code and developed an entire framework X in the code necessary for my research. Developing framework X took at least four months. After I finished the project and was put on a new project, he initiated collaborations with people from other labs using framework X (not including me). Without framework X this research would not have been possible (unless he would have implemented it himself).

Now I noticed that they are writing up an article about their collaboration. So far I have not been contacted as a co-author. My question is: should I claim co-authorship based on the fact that I developed framework X, which formed an important and necessary part of the code to obtain their results.

Optimist
  • 389
  • 3
  • 8
  • 43
    Presumably if the development of framework X is a meaningful intellectual contribution, you published a paper in which you describe it? So, he can just cite your paper. – ff524 Nov 28 '16 at 10:16
  • Framework X itself is not a scientific contribution so I will not publish it. However, I do think it is a intellectual contribution, but I will need to look up the exact definiton. – Optimist Nov 28 '16 at 10:26
  • 6
    It doesn't sound like you've earned authorship, then. Didn't you write a paper based on the research you yourself did with Framework X? Did you describe the framework in that paper? – ff524 Nov 28 '16 at 10:28
  • What did the framework do? Did it supply an essential algorithm? Did it present the problem into a fundamentally new way (and did you come up with this new way of looking at the problem)? Or was it as essential as the "framework" of the building they were working in: essential, but not intellectually very relevant. – dimpol Nov 28 '16 at 10:30
  • 1
    I provided an essential algorithm. It did not present the problem in a new way. It is more like a building block necessary to do the actual research. Based on comments I feel that I indeed do not deserve authorship, but this really left a bad taste as I feel taken advantage of. – Optimist Nov 28 '16 at 10:36
  • 2
    Note that (at least in the case of computer science), many conferences have specialized tracks for work focusing on tools/implementation. Since submissions for these tracks are not expected to have a scientific contribution, they might be suitables places to present your framework. – lighthouse keeper Nov 28 '16 at 10:43
  • @Optimist if you make Framework X open source, then it is a scientific contribution and can, indeed, be published. – Ander Biguri Nov 28 '16 at 12:35
  • 17
    If you compile your framework using GCC or if your framework is written in Python, do you include the people who developed GCC or Python as co-authors on any of your papers? – mikeazo Nov 28 '16 at 13:06
  • 35
    Authorship no, a citation or acknowledgment yes – smci Nov 28 '16 at 13:44
  • @Optimist: In order that people know the value of your framework, you can of course give a list of publications that used your framework. Just like the authors of Python could list important projects using Python. Or a typewriter manufacturer could proudly list some bestsellers written with their typewriters. – gnasher729 Nov 28 '16 at 13:53
  • 3
    @AnderBiguri Open- or closed-source is orthogonal to whether it is a scientific contribution. A scientific contribution is a knowledge advancement, such as a novel algorithm that solves a specific problem better than previous algorithms. A new open-source implementation of a known algorithm is not a scientific contribution. – lighthouse keeper Nov 28 '16 at 14:07
  • @lighthousekeeper I guess I disagree. Tools to make scientific advancement easier for others fall, in my opinion, inside scientific advancement. If someone invented a new microscope, that would surely be considered an advancement. I think the same applies to software. Truth is there are tens (hundreds probably) of software packages that have been published in scientific journals, mine included. – Ander Biguri Nov 28 '16 at 14:11
  • @Ander Biguri In my view, if someone invented a new microscope, that would fall in the same category as describing a novel algorithm, which is a scientific advancement. In turn, making that microscope available to a larger community of scientists (like OS packages) is not an advancement in itself. I have also published articles on tools I built, but I would not claim them as scientific contributions - except for the ones that included some form of additional validation, which is a form of knowledge advancement. – lighthouse keeper Nov 28 '16 at 14:20
  • @lighthousekeeper ahh I guess it comes down to the philosophical view. I sometimes think that its is even a bigger scientific contribution to the community to give them tools that they did not have before than inventing a new way and not giving software to do it. I understand your point, I just see it another way around. For the shake of this SE post: if the author would open source it and publish a paper about it, his/her problem would be solved, which was my point – Ander Biguri Nov 28 '16 at 14:41
  • 1
    @Ander Biguri I entirely agree that's an important form of contribution and it deserves much greater recognition than it often gets. I'm just not sure if labelling it as a scientific contribution is the best way to achieve that, but that's indeed philosophical. To the more concrete side, at least in my research area, open- or closed-source is not a main assessment criterion for tool-focused papers (although it maybe should be). – lighthouse keeper Nov 28 '16 at 14:49
  • @lighthousekeeper well, you can not publish scientifically a closed source toolbox. But yes, it has to be more than just "open source" – Ander Biguri Nov 28 '16 at 14:52
  • "Can I claim authorship on work I wasn't personally involved in" No. – xLeitix Nov 28 '16 at 18:44
  • Do you think the company/author od the language your wrote the code in wants the same treatment as you? – PmanAce Nov 29 '16 at 15:03

6 Answers6

37

Disclaimer: This is somewhat field-dependent. I'm answering based on what I believe to be the current situation in computer science / related fields.

1) Can you claim authorship given the way things have evolved? Probably not.

2) Would it have been a good idea for him to find a way to include you in the research collaboration so that you could get something out of your work? Yes.

The area of academia in which I work (*) is structured so as to value ideas over implementation. In practice, that means that it's not uncommon for people who write large amounts of code for a project to get little credit for it, which seems inherently unjust.

The way sensible people usually handle this is to make sure that people who do a lot of implementation work are given an opportunity to contribute scientifically over and above that - that allows them to justify their authorship in scientific terms, regardless of the fact that they also made significant non-scientific contributions to the overall project.

It's certainly possible to take advantage of implementers by not doing this, but people who do that tend to find themselves doing the implementation work themselves.

Bottom line: He should have had the common sense to include you in the scientific collaboration, and if he didn't have that common sense then his supervisor should have suggested it.

As an aside, I would view the situation somewhat differently if you already had a publication as a result of your work. In that case, you've already got the credit for working on the framework, and there's less of a moral obligation to include you in the new collaboration: instead, they can simply cite you. I'm certainly not of the view that writing a framework entitles you to be involved in every paper that is ever built on top of that framework - indeed, I've written frameworks that other people have continued to use for later research, and I wouldn't expect to be involved with the papers in question.

(*) As pointed out in the comments, this may not be the case in other fields.

Stuart Golodetz
  • 5,023
  • 1
  • 24
  • 23
  • I can see that this makes sense. I think I learned a good lesson here for future collaborations. I feel that he took advantage of me since he switched his research into my direction from the moment framework X allowed him too. – Optimist Nov 28 '16 at 10:42
  • 1
    I think a usual way to handle the situation of "code written by X for project A reused by Y in project B" is to cite the paper of the person who wrote all that code. E.g., Y would write: "We use the Python implementation of the Fundamental Theorem of Basket Weaving written by X [1]." where [1] is the paper written by X for project A. – ff524 Nov 28 '16 at 10:43
  • 1
    @ff524: Yes, assuming it resulted in an actual publication (sometimes it doesn't). – Stuart Golodetz Nov 28 '16 at 10:44
  • 7
    @Optimist: The bottom line is, there's an informal fairness code to be followed with things like this, and unfortunately not everyone follows it. My view is that you learn by experience which people don't follow it, and you avoid working with them. – Stuart Golodetz Nov 28 '16 at 10:47
  • 1
    It sounds like your answer is more about the case where X writes code for Project A, and then isn't given a chance to contribute more to Project A so as to earn authorship. Not necessarily the case where it's a different project, and X can still publish on A if he wants to get credit for his work. (The OP hasn't clarified whether he has a publication related to the original project or not.) – ff524 Nov 28 '16 at 10:47
  • @ff524: My answer's more about the case where X writes code for no particular research project (i.e. to contribute to the overall framework of the group), and then projects are built on top of it without giving X any credit for that. But yes, if a publication came out of the original work, then it can just be cited. – Stuart Golodetz Nov 28 '16 at 10:49
  • 1
    Ah, I see. I had the impression that the OP developed it to advance his own research ("developed an entire framework X in code necessary for my research") not as a generic group contribution. It seems odd to assume the PhD student is responsible here and accuse him of lacking "common sense"; maybe the student believed (and had reason to believe) that the OP would have published his work by now, and that he could then cite it? (Maybe you could explore some of these subtleties in your answer.) – ff524 Nov 28 '16 at 10:51
  • 2
    @StuartGolodetz your answer is exactly adresses my case. I have learned from this and will avoid working with this person in the future. – Optimist Nov 28 '16 at 10:55
  • "Academia is structured so as to value ideas over implementation". While this is true, he should still at least be one of the authors of the paper, just not the first one. He did the work, so he should be an author. Simple as that. – Louic Nov 28 '16 at 11:08
  • @LouicVermeer: I didn't say it should be structured that way, I said it is. As it happens, I think there's a serious problem there that needs to be fixed. In the meantime, one way of mitigating it whilst remaining within the current rules is to make sure that implementers are given the opportunity to contribute to the science and not just the code. – Stuart Golodetz Nov 28 '16 at 11:57
  • @StuartGolodetz: In my field (biophysics), everyone who contributes will be an author of the paper. Even if they only spent the whole day pushing a button and changing a sample to do a measurement, simply because they invested their time they are an author of the paper. Maybe it is different in your field or country but you can certainly not generalise to all of "academia". – Louic Nov 28 '16 at 12:42
  • @LouicVermeer: Fair enough, I stand corrected - let me change my answer to make it clear. – Stuart Golodetz Nov 28 '16 at 13:56
  • 1
    @Optimist I don't see why you shouldn't be working with this person anymore. After all, he's a PhD and he likely wasn't aware of the ethics expectations. Second point: can you publish your framework so the collaborators could cite you? – svavil Nov 28 '16 at 17:21
  • By writing "the current situation in computer science / information engineering" it looks like these fields are similar or even the same, while science and engineering are actually very different fields. – Andrea Lazzarotto Nov 29 '16 at 19:03
  • @AndreaLazzarotto: Sorry, perhaps I'm not wording it as clearly as I might (the context is that I'm a computer science DPhil who's now working in a computer vision group in information engineering - the two fields overlap considerably, to the extent that people often submit to the same conferences, and the fields seem to me to work in quite similar ways). I'm not implying that science and engineering per se are the same thing; however, information engineering seems to me a lot like applied computer science. How could I word it that would make it sound clearer to you? – Stuart Golodetz Nov 29 '16 at 22:54
  • @StuartGolodetz if that is the meaning you want to convey, I honestly don't know. :) This is because I do not agree with your view. I have a MSc in computer science and a bunch of friends who study or have studied information engineering (some of them even switched from one to the other) and I cannot agree to the statement that computer science and information engineering are similar. At least in my country (Italy). :) They share programming, a little bit of signals-related stuff and that's it. I'm a computer scientist, not an engineer (and viceversa for my friends). – Andrea Lazzarotto Nov 30 '16 at 00:01
  • @AndreaLazzarotto: I guess maybe I'm just working in the overlapping bit and that's why I think they're similar :) What I do as an "information engineer" feels a whole lot like applied computer science to me. (Indeed, what I do is done in the applied half of computer science departments in some universities.) Perhaps I shouldn't generalise to information engineering as a whole. – Stuart Golodetz Nov 30 '16 at 00:36
  • For what it's worth, the way both fields are practised here, there's such an overlap that people keep mooting the idea of putting the information engineers and applied computer scientists in a single building. YMMV. – Stuart Golodetz Nov 30 '16 at 00:39
16

Approaching this question from a psychological rather than legalistic perspective, I would argue similarly to vadim123 that the question you should ask is not "should I claim co-authorship?" When you feel a need to "claim" anything, you are putting yourself in an adversarial position relative to the student and other collaborators. This is likely to antagonize them and put them on the defensive with the perception that you are coming to stake a claim to a piece of property they have ownership of. Given the marginal nature of your claim, this zero-sum type of logic would mean that your chances of success would be very slim and you would run a substantial risk of hurting your reputation and/or souring your working relationship with the paper's authors and other colleagues.

On the other hand, it's important to keep in mind that this is not in fact a zero-sum situation. While I don't think you should be talking about "claiming" authorship, there is nothing to stop you from suggesting that you may become an author to the work by participating in its writing and including the framework you wrote as one of the topics discussed in the paper. The key is to do this in a diplomatic, sensitive way that appeals to the authors' self-interest and points out various ways in which it might very well be to their benefit to add you as a coauthor. It's also important that your offer makes it clear that you don't intend to force the issue and consider a decision not to include you as a coauthor to be a reasonable and legitimate one. Above all, avoid any use of accusatory, entitled-sounding or negatively charged words (such as "claim" or "deserve").

Here's an example (perhaps slightly exaggerated for dramatic effect) of how an email to the student bringing up the issue might work:*

*credit to @ff524 for the fictional - I assume - basket weaving research theme ;-)

Dear Joe,

I heard that you and [insert name of group/collaborators] are working on a paper on your new results about basket weaving. Congratulations, I'm glad something came out of that project. I hope you manage to publish it somewhere good.

I thought I'd run something by you in connection with that. As you know, I spent several months working quite hard on the Basketweaving Framework that you used in that analysis. While I don't necessarily think that that should entitle me to coauthorship in your paper (and if you don't want to offer me coauthorship, I understand and think that's completely reasonable), it did occur to me that if you were to offer to let me take part in writing the paper and become a coauthor you could get several benefits that might not have occurred to you:

First, we could add a section about the basket weaving framework itself, which I created. I don't know if you wanted to mention it, but the new content could improve the paper and make it appeal to a wider audience, and possibly be accepted to a more prestigious conference, etc.

Second, another way that I could help is that I am a really good writer and put a lot of work into each paper I write - you can look at my existing papers and see how polished they are, with great figures etc. - and I can assure you that I am willing to work very hard to ensure that this paper has the same level of polish that would help the quality of the work shine through. Did I mention that my most recent paper was accepted to the International Congress Of Basketweavers? ;-)

Third, I've found in my career so far that while having coauthors does dilute a bit the credit you get for papers you write, it does so much less than you might think, and this effect is usually greatly outweighed by the fact that a coauthor is essentially a free spokesperson working for you, who goes around giving talks about your joint work at various places, and who can often say nice things about you in those talks that you couldn't easily get away with saying about yourself. Maybe it's cynical of me to be thinking in this way, but that's just one of the quirks of academia and how it operates and it's just as well to be aware of it.

As I said I'm happy with whatever decision you and the group end up making, just thought you might want to think about this.

Community
  • 1
Dan Romik
  • 189,176
  • 42
  • 427
  • 636
6

Authorship is much like right-of-way on the roads, in that it can be given but not taken. The answer to any question that begins with "Can I claim authorship" is no; instead one should ask "Should I receive authorship". Of course, both drivers and authors sometimes make mistakes (and can be criticized or punished for those mistakes), and sometimes people get hurt.

Such difficulties can be avoided by simply discussing, at the beginning of any research collaboration, what the possible authorship outcomes might be from the results. With an agreement in place, a potential he-said-she-said situation disappears, and a withheld coauthorship becomes a clear offense.

As the other answers and comments indicate, the specifics of the OP's situation is highly contextually dependent. OP should discuss how to proceed with senior members of his or her department, who will have more useful advice than strangers on the internet.

vadim123
  • 4,192
  • 17
  • 22
4

Isaac Newton is famous for saying:

If I have seen further it is by standing on the shoulders of giants.

and hopefully that's what the Ph.D. candidate has done.

The most important advice I would give you is talk to them, be forthright but not aggressive or accusative (certainly not to begin with). Use non-definitive language, i.e. explain how you see what has transpired, say that it seems to you that your contribution in writing the code may merit authorship - yes, don't be shy about that - and that you would like for them to let you know how the paper is going to look like for you to be able to understand whether or not that is the case. Don't claim they have done something wrong and don't talk about their motives. Make it seem like they would merely be putting your mind at ease.

About the actual question of whether you should be credited as an author or not:

  • If the paper doesn't claim (explicitly or implicitly) that they wrote the "framework", or the significant part of the code that you wrote; and if it doesn't present that code/framework as its novelty - then No, you can't and shouldn't claim authorship.

  • If those conditions are not met then Maybe, depends on the specifics. This is why my suggestion is to try to get more information; it might be the case that you can't expect more than a thank you note, it might be that you should be an author of that paper, and it might just be that they just refer to the framework as an interesting endeavor in itself which should be described perhaps in another paper than the one they're writing.

einpoklum
  • 39,047
  • 6
  • 75
  • 192
0

Even if he uses your framework, he wrote the code that uses it. So he is not required to let you claim authorship of it. However, you may be able to politely ask him to credit you (or similar).

Dev
  • 111
  • Please read that the OP writes "I made significant contributions to his code". – Nobody Nov 29 '16 at 06:24
  • @scaaahu The OP only asserts that on the basis of having written the framework, which is not what Dev seems to be discussing. – Fomite Nov 29 '16 at 07:18
  • @Fomite I'll emphasize "I made significant contributions to his code and developed an entire framework X". Dev says "Even if he uses your framework, he wrote the code". It seems to me Dev thinks the PhD student(he) wrote the code without the OP's involvements. – Nobody Nov 29 '16 at 07:29
  • @scaaahu I was more going off the second paragraph, where the OP seems to advance that the primary contribution was the framework itself. It's enough to be a little ambiguous. – Fomite Nov 29 '16 at 07:31
  • 1
    @Fomite I agree the question is a little ambiguous. But, I think this answer is too ambiguous to be a good quality answer. – Nobody Nov 29 '16 at 07:33
0

I completely agree with the accepted answer, assuming the contribution of the paper requires a reference to the output of the software.

In Computer Science (and in connected applied fields, such as Bioinformatics), the general tendency is to consider software as a "bona fide" scientific contribution, analogous in term of technicality and required rigor to a mathematical proof (the writer of which would certainly not be denied authorship in a Maths paper). So I would say the sole contribution of an unpublished critical component of the research in itself qualifies the OP for co-authorship of that specific study. Field-specific conventions (first/last author, stars and daggers, dedicated section...) may be able to clarify the contribution of each author.

Moreover, in computational fields (as in "Using computers to analyze genomic/astronomic/ecological... data"), a reference/url to the software used for the experiments would typically be required by journals for the sake of reproducibility. Since you did not publish the implementation, then the reviewer may not be able to reproduce the claimed result. So, unless you plan to publish your implementation at a later time, the easiest option would be for the authors of the study to include you and mention this novel piece of software.

Yann Ponty
  • 36
  • 4