14
$\begingroup$

I am writing two separate paper that are closely related. When I try to submit to arXiv, is it possible for each paper to refer to the other paper with an arXiv link, rather then putting a newer version of one of them just to replace the arXiv link in the reference? Since I heard from others that too many frequent update to a new paper is not preferable.

So, what is the usual practice in case such situation happens?

$\endgroup$
11
  • 2
    $\begingroup$ You have a window of opportunity of a few hours when you can still edit your submission before it is put into arXiv. During it you will know the arXiv id of tha article. Hence you can submit both of them,get their ID's and then insert them. $\endgroup$ Oct 15, 2011 at 4:27
  • 17
    $\begingroup$ Torsten's approach used to work, but more recently the arXiv has switched to using "temporary submission identifiers" and you don't learn the arXiv ID until the article actually appears on the arXiv, at which point it is too late to replace it without generating a new version. $\endgroup$
    – Henry Cohn
    Oct 15, 2011 at 6:08
  • 3
    $\begingroup$ Putting just one new version (or even two) is not a big deal. The "warning" on arXiv is to avoid having "Version 12" or something that high, which unfortunately does happen. And what you suggest (replacing a version just to add an arXiv link to another paper) does happen very frequently. $\endgroup$ Oct 15, 2011 at 12:27
  • 9
    $\begingroup$ I've decided to unilaterally reopen the question, as I can offer a definitive answer, in the form of the reply from the arxiv administrators when I asked about this issue some time ago. $\endgroup$ Oct 15, 2011 at 18:16
  • 4
    $\begingroup$ (Also, I think this is a reasonable question.) $\endgroup$ Oct 15, 2011 at 18:19

1 Answer 1

20
$\begingroup$

I asked the arxiv administrators about this last year, and received the following reply.


Dear Scott,

Thank you for your feedback. At this time we have no plans to change the way the submission ID and the final arXiv ID are created. We developed the new submission system to be more flexible for our users who need an area to work on their submission and conduct any updates for formatting/layout/typos without being constrained on time. Hence, we can not issue an arXiv ID until the submission has been announced.

We are looking into ways to make a submission available to all the co-authors but can not provide a time frame when this feature will be added. We have received feedback that validates the need to be able to share a submission before it is announced and we are working on a solution.

-- arXiv admin

On 07/07/2010 11:38 AM, Scott Morrison wrote:

Keyword: "submit/0016922"

Dear arxiv admins,

I'm concerned about two ways in which the new submission system is inferior to the old one (mostly, though, it's great!)

First, we used to be told what the arxiv identifier would be. Now we're just given a temporary identifier, like submit/0016922. This is a problem for me, as I always used to include the arxiv identifier in my paper (usually as a footnote, e.g. "This paper is available at http://arxiv.org/... and at ..."). It will actually be a real problem, presently: my coauthors and I have 3 related papers that we want to put on the arxiv simultaneously. We'd like to be able to refer from one paper to the others, but it seems now there's no way to do this without submitting a second version the next day, because we can't discover what the identifiers will be.

Second, it's somewhat annoying that there's no way for a coauthor to see the submitted paper before it is announced. In the past, we would always have someone other than the submitter carefully check the produced PDF, abstract, etc., but it's now much harder to do this.

I'm not sure if it's possible for you to remedy either of these problems, but I would encourage you to try!

scott

$\endgroup$
5
  • 1
    $\begingroup$ Presumably, your coauthors have checked the final product before you submit to the arXiv. Does it ever happen that the arXiv's compiling into PDF differs in any serious way from what you see when you compile the paper yourself before submission? This would be news to me. $\endgroup$ Oct 15, 2011 at 18:24
  • 4
    $\begingroup$ @Spiro, yes, definitely. In particular, I remember an instance when helping someone submit their paper, that somehow all the figures got appended at the end of the paper on separate pages, as well as where they were meant to be. I forget the mechanism (it was somehow our fault, rearranging some directory structure), but we barely noticed it. I think I also recall some problem with included postscript, with line sizes being mysteriously changed. I've also made mistakes in preparing the "plain text" abstract for the metadata from the LaTeX sources, which coauthors have caught in time. $\endgroup$ Oct 15, 2011 at 22:56
  • 4
    $\begingroup$ There are a couple of ways the arXiv PDF can differ from what you expect. One is that some LaTeX packages exist in several different versions; usually, they don't differ in the output except for extending the functionality or fixing bugs, but occasionally they produce visible changes. (This happened with the AMS document classes, which was a real hassle for the arXiv.) The other issue is that some packages, such as microtype, produce different output under pdflatex. By default the arXiv goes from dvi to ps to pdf, so if you need pdflatex you have to set the \pdfoutput flag in the TeX file. $\endgroup$
    – Henry Cohn
    Oct 16, 2011 at 5:04
  • 1
    $\begingroup$ I believe that the figure appending issue occurs when you are using regular LaTeX (not pdflatex) and you have figures which you uploaded to the arXiv but don't use in the paper. To get an extra copy of all of the figures, I think the most likely cause is having two copies of all of the figures in separate directories. $\endgroup$ Oct 16, 2011 at 7:04
  • $\begingroup$ I guess my papers don't have enough fancy TeX in them to cause such problems. I've had some figures, though. All seemed to work fine for me. But I'll definitely check it carefully next time I submit. $\endgroup$ Oct 16, 2011 at 12:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.