8

I've recently submitted a paper to a conference requiring anonymous submission, meaning that the name of the authors should not be included, and no obvious self-reference should be made. The problem is that I add additional information relevant to the submission, such as some proofs and some code working on an example, calculating some numerical data (that could be found on the paper, but requiring some effort).

For a non anonymous submission, I would have published a technical report with all the proofs, cited the report in the submission, hosted the code on my website, and put a link to it. However, since the technical report would not have been anonymous and since my website is obviously linking to me, it was not possible.

Concerning the proofs, I managed to put them in the appendix, explaining that they were only intended for the reviewing process, but I could not find any solution for the code. The conference submission site (easychair) did not provide any way to input additional data. What's the best way to deal with this kind of situations?

9

There are three criteria: the solution must obscure your identity, it must not allow you to learn anything about the reviewers, and the second criterion must be common knowledge.

Any attempt to distribute files yourself through the web will fail to satisfy these conditions. For example, if you use a website you own (under a pseudonym, say), then you can inspect the logs. Even if you have arranged to use a service that will not provide any logs or allow you to tag pages or link to outside services, the reviewers may not notice that you have done so, and they may not even be willing to believe that you are not doing something tricky.

So this is a problem you cannot solve yourself. You should get in touch with the program committee chair and ask whether the committee could make this data available to the reviewers. As long as it goes through them, it doesn't matter how they provide it.

  • I agree, the conference should take care of that, I'm actually surprised that conferences requiring anonymity don't offer such a service by default. Do you know any conference that does? – user102 May 8 '12 at 15:05
  • "if you use a website you own... then you can inspect the logs" - this seems like an exaggerated limitation to put on an author. I never encountered a situation where a paper was rejected based on this concern. If this is indeed a concern, it makes more sense to instruct the reviewers to use a proxy that protects their identity. – Erel Segal-Halevi Oct 31 '13 at 18:03
  • 1
    @ErelSegalHalevi: See my comment to aeismail's answer for a real-life example, although I doubt this is common. To me, it seems easier to have the editor host the files than to try to get reviewers to access them through a proxy, but either way could work. – Anonymous Mathematician Oct 31 '13 at 21:54
5

If you need to distribute stuff anonymously via the web, as far as code goes, you could create a pseudonymous github or Google account for a "project," and then make it accessible during the review period. Once the review period is finished, you could delete the site, if intellectual property issues are a concern.

  • 1
    ...unless the reviewers also anonymize themselves via proxies and whatnot. – JeffE May 8 '12 at 13:44
  • 1
    @JeffE I agree with you, the reviewers can protect their anonymity themselves, I'm not sure though how many could technically do it :) (I know several researchers who would have no idea of how to set up a proxy!). – user102 May 8 '12 at 15:04
  • 1
    For 300 lines, you could just include the code on one page of your appendix, and let the reviewers copy and paste from their pdf viewer. – JeffE May 8 '12 at 15:08
  • 4
    @JeffE: The problem is reviewers who don't even realize their anonymity is threatened. It's rare for this to occur, but it has happened. I know of one case involving a journal submission. One referee report complained that the simulation results weren't convincing, so the authors offered additional data for the referee on their web site, and then identified him based on the logs and complained to the editor about why they believed he was biased. The most shocking thing was that the authors told the editor what they had done. I'd guess that it happens much more often secretly. – Anonymous Mathematician May 8 '12 at 15:45
  • 2
    @AnonymousMathematician: This is clearly something that needs to be managed by the conference organizers, then. If they want to insist on total anonymity, then they need to provide the resources to make that happen. Perhaps the code is provided to them, and they host it on a website. Or they create the repository, so that the file provider doesn't know the URL. – aeismail May 8 '12 at 17:30

Your Answer

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