I'm finding that this is a problem for me.

All of my papers have been solo efforts so, naturally, nobody else is really invested in them before I submit to journals. After I submit a paper, I either notice mistakes or cannot resist making significant stylistic changes. Then I send a revision to the editor to forward to the referee. This has happened multiple times, sometimes more than once for a particular paper.

I would like to hear from other researchers:

(a) Do you have similar tendencies? (b) As a referee, how frustrating is it to get these revisions? On one hand you're getting an improved version (in the eyes of the author), but on the other hand some of the referee's work may be wasted.

  • 8
    I'd advise to, after finishing the paper, leave it for a week or so - don't submit it, don't read it, don't do anything with it. Take a break, focus on other taska etc. After this week, read the paper like it's not yours, like you've seen it for the first time. Note what you don't like about it, point out errors/mistakes; act like a referee for yourself.
    – user68958
    Feb 22, 2018 at 7:12
  • Even for a solo effort, getting a friend (or one of your students, who might learn something from it) to proof-read the paper before submission will catch a lot of these. Not all, of course. This can take place in the waiting period @corey979 advises
    – Chris H
    Feb 22, 2018 at 16:46

4 Answers 4


(a) No, I don't have similar tendencies. When I submit a paper for publication, I'm so relieved to have finished it (usually much later than I had planned) that I usually ignore it until I get the referee's report. I usually have plenty of other tasks that need my attention because I had postponed them while finishing the paper. (b) As a referee, I don't mind getting an improved version of the paper provided it comes with clear information about where the changes are. I don't want to check the part of the paper that I've already read (and my notes on that part) against the new version just to find that nothing has changed in that part.

  • (a) But what if a reader points out errors? Feb 22, 2018 at 3:20
  • 3
    @darijgrinberg Good point. Unfortunately, my errors were pointed out only after publication, but if I were informed of an error while the paper is being refereed, my reaction would depend on the severity of the error. If the error is trivial, I'd probably take care of it later, when I prepare the final version of the paper. If the error is nontrivial but I know how to correct it, I"d inform the editor and prepare a corrected version of the paper as quickly as possible. If I don't know how to correct it, I'd withdraw the paper. (In all cases, I'd thank the reader.) Feb 22, 2018 at 3:52
  • 2
    @SolarMike: Readers comment on arXiv preprints a lot. Feb 22, 2018 at 6:03
  • 1
    @darijgrinberg arXiv seems to be fairly limited in fields offered : so is it representative of the bulk of papers published worldwide or only a fraction? A quick look and I did not see CFD Computational Fluid Dynamics for example.
    – Solar Mike
    Feb 22, 2018 at 6:41
  • 2
    @SolarMike Note that this question is tagged "mathematics", which is well represented on arxiv.
    – Arnaud D.
    Feb 22, 2018 at 10:41

Please don't do this. Especially multiple times for the same paper. It wastes the referees' time and I'd be very annoyed by it. Don't keep moving the goalposts.

If you find an actual mistake that requires significant changes to the paper, contact the editor. For anything else, wait until the referees have finished their job and then fix it. That includes any kind of stylistic change and any minor mistakes that don't have significant effects, unless you think it would take the referees more than a short time to figure out what's going on. (For example, if your paper claims that 2+2=5, that's not going to cause much confusion to the referees so you don't need to send a new version. However, if you make some long argument based on the "fact" that 2+2=5 and that argument requires significant changes to cope with the correction, you should talk to the editor.)


IMHO I think you should let the referee read and comment on the original version, then you can include the revisions asked for as well as the errors you wish to correct as well as your stylistic changes. This is what my supervisor did with me when we had errors we wanted to correct.

This minimises the work of the referee on your paper, also you should perhaps be more critical / thorough on your original before submitting it.


(a) I occasionally edit my preprints after submitting them to a journal, although not nearly as often as before that. As you guessed, this might confuse referees, and most likely they'll want to know what exactly has been updated. When I'm the sole author, I solve this problem by complete transparency: all my changes go on a public git repository, and I tell the referee which commits have happened between the previous and current revisions. (Example git repository -- but this one is nowhere near getting submitted yet. Note that using git is extremely easy if you're the only author and are only using it to keep track of your own changes -- in this case, you don't even need the commandline; everything can be done by point & click on github.)

(b) Yes, as a referee I am often frustrated when I don't know what exactly has been changed (PDF files are painful to compare, if the changes aren't completely minor). Authors should strive to make changes as transparent as possible; but this doesn't mean that they should avoid making changes.

  • 1
    But is it a good idea to share publicly before the paper is accepted ? Would you accept the same as a supervisor or project investigator ?
    – Neel Basu
    Feb 22, 2018 at 15:59
  • @NeelBasu: It's utterly standard in mathematics (this question is tagged mathematics). Feb 22, 2018 at 17:36

You must log in to answer this question.

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