30

I've been using git for a while and now GitHub. I'm starting to write a master's thesis. Would it be a good idea to upload all the files to online public repositories? Does anyone have any experience with this?

EDIT: Thanks for all the answers. I consulted with my adviser about this, and he said that latex sources were ok, as the document will be open access anyway. My source code however was not, primarily because the collaboration already has a protected wiki for such things.

  • 5
    Github offers private repos for academic users. Also, see this Tex SE question for a lot of advice about using LaTeX for writing a thesis. – eykanal Feb 26 '12 at 3:28
  • 1
    @eykanal, thanks for the link but unfortunately github is not currently offering free private repositories for academic users... they say they will offer more "soon" though. bitbucket.org offers free private repositories for all though – Michael Bishop Mar 16 '12 at 23:53
  • Now github does, I have one and I'm very happy with it. :) – crsh Oct 12 '12 at 12:02
  • @crash they have "educational" discounts, but these are not available for academic use in general - unless I am missing something github.com/edu. – David LeBauer Dec 22 '12 at 4:47
19

I'm told for your field that the answer might be yes, because you're slightly less threatened in terms of being scooped than most, with physics (apparently) going largely by who submits first.

I would be extremely cautious publishing the source of my thesis in its entirety on a public repository. I'm all for repeatable science and open access, but the public should only have access to your code when this condition is met:

You have no further use of the exclusive access to your code. All the questions you've programmed have been answered, the papers and presentations that emerge from them are in press, and at this point, it is a question of reproducible research.

Once that's true, sure. Until that point? You're running the profound risk of your research being stolen.

  • 5
    You can also do it when your research is so amazing that it opens up whole new fields that you couldn't possibly explore in a single lifetime. – David Ketcheson Feb 25 '12 at 18:04
  • 2
    Wait, what? Doesn't posting your work on github give it a time stamp? (Although I suppose if you're really worried, you could post it to the ArXiv at the same time.) – JeffE Feb 25 '12 at 19:49
  • @JeffE It gives you a time stamp - but it doesn't necessarily mean some fields will regard that time stamp as "first in". It's not exactly hard to modify someone's code enough, or obfuscate by not releasing your own code, to make the scooping/plagiarism risk a real one. Beyond the risk/whether or not you'd come out ahead, in my mind never having to deal with an IP theft scandal (even as the wronged party) as a graduate student exceeds the marginal utility of your stuff being open. – Fomite Feb 26 '12 at 3:53
  • 3
    @EpiGrad: I simply can't imagine anyone getting away with that kind of robbery in the face of easily available evidence. I've always thought of public, easily-verifiable timestamps as a defense against scooping. I've seen (a few) papers rejected because the authors didn't cite a month-old independent ArXiv preprint with similar results. I've also seen authors scooped because they didn't publicize their partial results. – JeffE Feb 26 '12 at 11:07
  • 3
    @EpiGrad: Again, I know of instances in my field where a student/postdoc was scooped because they didn't publicize their preliminary results. (What do people in your field do when faced with evidence of plagiarism?) – JeffE Feb 27 '12 at 14:33
18

The answer is in two parts:

  • Yes, it's a very good idea to use an online repository with a versioning system to write your Master thesis. It offers a nice automatic backup, you can easily sync from different locations (office, home), and (this is mostly true for papers rather than a thesis) you can easily collaborate with people outside of your university (i.e. who wouldn't have an account on your university server.

  • No, it's not a good idea at to make it public. Plagiarism is real, online versioning system do not offer any real protection, and you don't want to make it too easy to copy, especially when you're not finished yet. I'm all for the open-access of finished documents, that you can put on arXiv, but in this case, there is a real timestamp, and your work can be indexed (for instance, in Computer Science, arXiv is indexed on DBLP).

For these reasons, I've been personally using for my papers BitBucket, that offers for free to academics their unlimited plan (unlimited public/private repositories, unlimited collaborators). In this way, I create a private repository for each of my paper, give the access rights to my co-author. It seems that GitHub also provides a way to have a private repositories for academics: https://github.com/edu, so you can keep using it for sharing your open-source code (for instance) and use it with a private repository for your thesis.

10

Talk to your adviser about this. Do not make anything that is unpublished publicly available without their knowledge. Any other people who might be co-authors on your papers might also have a say.

Having said that, a private github repository is a great idea (I used one myself for my PhD dissertation & papers).

8

Code - yes, it is a good idea.

PDF (perhaps with LaTeX source) - you can (why not? but rather as a place to store it and point to).

However (for finished), there are more dedicated places, e.g.:

  • opentheis - repository of theses,
  • Mendeley - reference manager allowing to upload your publications (including theses),
  • arXiv - base of preprints (but only for theses in English and from a list of topics; there you need to put source as well),
  • etc.

When you intend to actually write your thesis in the public - well, it is mostly up to you ( i.e. if you fill comfortable with it).

  • not pdf, but the directory of latex source – Approximist Feb 25 '12 at 1:25
  • @Approximist Clarified. To be short: why not? – Piotr Migdal Feb 25 '12 at 1:34
  • 1
    By default, ArXiv requests latex source and generates its own pdf output; latex source is available for most ArXiv papers. See arxiv.org/abs/1202.2638 . – JeffE Feb 25 '12 at 19:50
5

You should be very careful. As usual, this depends on your field. Mine is mathematics and it is not uncommon for a graduate student's thesis question to be solved much more quickly by someone far more advanced. If your thesis is publicly available then you increase the chances of this. This is more relevant for those working on original research, which is not always the case for a masters thesis, but keep this in mind.

4

I don't think there is a universal right answer to this. Some factors to consider, during the thesis preparation stage:

  • Are there intellectual property issues to consider in your thesis work? If so, that would mitigate against having a public repository.

  • Does your school or institution have rules against such behavior?

  • Does your school offer its own service, which may offer better security for your data?

After the thesis is completed:

  • Do you hold intellectual property rights to the original materials, so that you can publish it to a repository?

  • Will publishing to a repository conflict with your ability to publish the work in journals (you'll need to check the journals in question)?

  • Does your institution offer an archival service that will allow such access?

4

Bitbucket offers unlimited private repositories for academic accounts that you can get by using your uni email id. I'm writing my PhD dissertation using Lyx and use Mercurial for version control, so I store my code in a private repo on bitbucket. I think version control is indispensable for large research projects that are completed over a long time frame, so using a repository would obviously fit right in.

But making your stuff public serves no useful purpose, neither for you, nor for the academic community that would be interested in your research. You make something public when it is ready to see the light of the day. Till then, you receive feedback from your supervisor and colleagues, and improve the quality of your work until it is good enough to be of use to others

Making half-assed research public is a crime against academia, can reflect poorly on you, and in any case conferences are the best venue for discussing work in progress and for getting your work time-stamped by a large gathering of people in your field who know each other, and this can prevent "scooping". However I must say I'm sceptical that such a thing even exists in academia, your research can only be done by you.

Journalists search for novelty, and can therefore be scooped. We academics, on the other hand, search for credibility, which only comes from knowing "your" idea inside out. If somebody else can do more credible work with your idea, then they get the Nobel, you get the citation.

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.