7

I am in process of uploading a paper for arXiv. I have looked over the paper for several weeks now and it seems that there are no major issues, however, the paper is 24 pages long so I have accepted the possibility of having a few (non-trivial) typos and grammar mistakes here and there.

My question is, what should you consider before uploading a revision to your already uploaded paper, should you spot a mistake?

I am only asking because it seems that the people I have came across in my field rarely upload a revision, whether or not there are mistakes in their paper. One, for example, contained a glaring mistake in an important part of a proof, but the mistake is still there; I can see it now. The paper has been there for a year now, it has gone through a formal revision at a conference (where the mistake was corrected). But why not simply correct their arXiv copy?

Should I wait until more feedback is received before uploading a new revision? If so, how long? Is it okay to upload the revised version immediately or is there some arXiv mechanism that prevents me from doing so?

Is there a catch for uploading revisions?

Improve this question

2 Answers 2

Reset to default
4

Another reason is that the arXiv itself somewhat discourages overly frequent updates. For instance, after the 5th version(?), all future updates are no longer publicized in the mailings, so you might end up with the first few trivial corrections seeing more publicity than the future ones. I am myself no stranger to the "I'll update the arXiv once I've changed something significant" behavior, even if the preprint is years old and it is questionable whether any significant changes are to come.

Improve this answer
1
  • 3
    Whatever. Most people find arXiv papers through Google searches and citations, not through email announcements. (I'm an arXiv moderator.) It's far better to have a few less people hear about the correction than to leave the error in place.
    – JeffE
    Commented Apr 8, 2017 at 18:58
2

Some journal policies can prevent people from uploading new versions of a paper based on feedback from their review process, at least for a certain time period. Some journals in some fields will even consider an arxiv version a "prior publication" of the paper, and as such won't bother to even review it.

Even if there's only a time delay, it's entirely possible that the author(s) simply forgot, as after 6+ months they are probably swamped in new projects, new semesters, etc. This can be compounded if the corrected version uses journal specific styles, and so would have to be edited to be suitable for the Arxiv.

Individual authors also use the Arxiv differently. Some of them use it simply as a way to announce their results once they've advanced far enough for a preprint. Such people are unlikely to bother with updating, especially if the paper gets accepted by the first journal it is submitted to. Others go further and view it as a paywall free way to maintain an accurate and up-to-date record of their work. These people are likely to update with corrected and published versions whenever possible, and may view doing so as an important (or even necessary) service to the research community. Many sit in between: they may update when significant changes in the results have been made, but won't bother with grammar and typo fixes alone.

In short: it's a matter of personal taste and preference. You don't want to submit an update for each typo you find, but you should at some point have a sense of "meaningful amounts of change has occurred, and I do not think I'll be making any more changes in the near future", at which point it would be appropriate to update if so desired.

Improve this answer
2
  • 1
    The first reason may well be non-existent. I have never heard of a journal halting the publication of a paper because an author uploaded a referee-corrected version onto the arXiv, let alone trying to go after the author in any other ways. (This is about "accepted manuscripts", not about the final versions formatted by the journal -- those even arXiv itself doesn't usually want.) This is not overly surprising, because the only people who could reasonably police these kinds of rules would be the editors, and those rarely have the incentives to harrass the research community they belong to.
    – darij grinberg
    Commented Apr 8, 2017 at 18:15
  • @darijgrinberg Agreed. Such policies do exist, but it's dubious about how enforceable they will really be, for the reasons you state. They are likely to discourage posting the update regardless, though, as the author similarly has no incentive to harass and antagonize the journals and editors they depend on. They might not try to penalize you for that paper, but possibly they will be less receptive to future submissions.
    – zibadawa timmy
    Commented Apr 10, 2017 at 13:50

You must log in to answer this question.

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