Am I missing something or is submitting from the ArXiv really that unfeasible?

Question

Some publishers, such as the APS, allow to submit papers from the ArXiv instead of from your computer. While this seemed somewhat reasonable to me at first glance, I noticed one problem with this that seems to render this feature pointless:

ArXiv submissions as well as replacements take some time, usually a few days, to be published. So, if you just submitted your files to the ArXiv, you would have to wait this time before you can submit to a journal. Instead of waiting, you could just take the files your just submitted to the ArXiv (or download your files via your ArXiv account) and submit them to the journal directly. This should take little additional time, as you do not need to prepare these files in any special way – if you would need to, submitting from ArXiv wouldn’t work anyway, as it would yield the same files.

Now, I can think of some exceptions where your files could already be published on the ArXiv. For example, you submitted a version to the ArXiv and to “internal” review at the same time and the internal review results in no changes whatsoever. Or you submitted a paper to the ArXiv and a journal at the same time and the latter desk-rejects your paper for being out of scope or similar; then you could submit to another journal from ArXiv. But these are exceptional cases: In most situations, you would decide that a paper is submit-worthy directly after you made some changes, which hence cannot already be published on the ArXiv.

Note that at least the APS does not seem to have a way to magically obtain the paper you just submitted to ArXiv as it requires an ArXiv number for submitting, which you only obtain after publication and hence after waiting.

As it would surprise me that time was wasted in implementing this feature if it really is as useless as I consider it to be, I am curious: What am I missing here? For example: Is there some common situation in which you would decide that a paper is ready for submission without any change? Or do you not have to wait for the ArXiv to publish your paper for some reason I am missing? Please answer only if you can address the above concerns.

Answer 1

I think you are not really missing anything at all, except that there is not compulsion to make the set-up as minimalist/efficient as you seem to argue it should be. That is, while you and others may not find this feature useful, it's probably harmless, no? And some programmer somewhere might've thought it would be interesting to implement this, and proposed it to whatever authority, it was approved, and so on.

But, yes, there are common situations for many people (in math, familiar to me) in which the submission to arXiv is essentially simultaneous with a submission to a journal, whether or not they make use of a literal submit-from-arXiv feature. So, yes, a paper might be considered ready without-change-from-the-version-submitted-to-arXiv. Why not? (Sure, one might realize things later, perhaps upon referee reports...)

As to whether one "has to wait until arXiv publishes..." Well, apparently, yes, since your paper won't have been processed. Perhaps someone who knows the inside workings of the software at arXiv (which surely could change from time to time) knows a trick to interact with the system prior to that, but I've never heard of anyone finding a need to do so, given the quick turnaround. But I can't help but wondering why a day-or-two delay would matter...

So I think you're right that it's a fairly useless feature, but it wouldn't be the first.

Answer 2

The submission and refereeing system used by the American Physical Society supports direct submission from the arXiv. However, I have found that the system does not work well enough for this feature actually to be useful. The APS system is not good enough at handling the metadata it extracts from the arXiv and from LaTeX source files to make submission from the arXiv worthwhile. (This has consistently been true of for my submissions, but perhaps other people using different style files and layout formats have had better results.) So when I want to submit a paper to an APS journal, I work directly from the source files on my own machine, not worrying about the arXiv submission feature.

In general, it takes about one business day from the submission of a paper to the arXiv to its becoming available for viewing. I have never worried about delays of a day or so in the submission process. Sometimes, I post papers on the arXiv first and don't submit to a journal for a while, because I want to see if I get any comments from other people in the field. A few times, this had led me to make changes to my manuscript before I sent a version to a journal. In other instances, I have submitted to the arXiv and a peer-reviewed journal at essentially the same time. However, even when I am anxious to get a paper into the reviewing queue quickly, I am not going to sweat about a one-day delay.