Eprint and DOI entries only for preprints and 'in press'

[nxg]

Keith T Smith, summarising an email exchange, on 2013 September 5:

7) arxiv numbers and dois are being added to every entry which includes them in the .bib . Whilst I can see why this is attractive to some, MNRAS style is not to include this information, except for

Yes. Elsevier style files do it and I think it's a great idea. DOIs seem to be absolutely perfect because they unambiguously point to the actual paper.

It would also be great if the MN class file enabled hyperref with approved styling for the links.

I agree that including these links would be helpful, but writing them out in full is not. To access them you need to be on a computer anyway, so a clickable link is fine. See next comment.

I'm not sure what ApJ do. What does "including but not writing mean"? You mean it somehow gets in the PDF?

Yes. ApJ include two clickable links as part of their reference lists: the journal name is clickable and goes to the DOI, whilst the volume and page numbers go to the entry on ADS. See e.g. http://iopscience.iop.org/0004-637X/775/1/1/pdf/0004-637X_775_1_1.pdf (paywall) as an example. I think this is a very good solution and something MNRAS should adopt. But until we do it in the final journal, we shouldn't do it in the latex preprints. hyperref for URLs and clickable citations is something I always do in my own papers, but the final version of MNRAS only has clickable URLs (and they're just black so it's not obvious). I'll definitely raise this at the next board meeting, which unfortunately isn't until March. I'd need to discuss it with the publishers first.

[nxg]

Re 'in press': how do we expect authors to indicate, in the .bib file, that an article is 'in press'? There's nothing at http://www.oxfordjournals.org/our_journals/mnras/for_authors/ which is explicit on the subject.

[timj]

I would also be interested. Historically I put "in press" for the volume and left the page blank. I wonder how this works with eprints though. These days most "in press" entries will have an arXiv reference but MNRAS seems to be indicating that an arXiv reference trumps all other information for a preprint so you would not actually have a MNRAS, in press at all but arXiv:1230.023 (eprint). It does seem strange to me to lose the journal name from the "in press" publication just because I also uploaded it to arXiv.

[timj]

To be pedantic, knowing a DOI when all you have is "in press" or arXiv is not actually possible.

[keithtsmith]

I don't think we have an actual rule about this. The publishers must deal with this all the time, but the editorial office always left it to them. I'll ping them an email to check current practice.

I think the best solution is something like:

• Authors et al., MNRAS, in press, arxiv:0123.4567

In that case, putting 'in press' in the volume field should work. Link goes on the arxiv ID as usual.

[timj]

I wish to correct my DOI comment above. I recently was on a paper accepted in A&C and that has been issued a DOI despite still being "in press".

[keithtsmith]

MNRAS does this too, although it's not well advertised. Once a paper is with the publishers, a DOI is created which contains the publisher's tracking number (e.g. stu237). The DOI continues to work once the paper is published.

[timj]

Using @keithtsmith example files I confirm that eprints don't turn up at all at the moment. In fact they stopped working in commit 3b36609bca08cfe6e155e3cfc9f5fc229f1818bf which is the one that jumped through hoops to try to hide the information on DOIs and arXiv entries once the article has been published. Looks like that logic needs a tweak. The DOI does get secretly added as a clickable link but eprint entries are completely ignored for journal articles: The "arXiv eprints" text is the link rather than the arXiv ID. The arXiv ID really should be present.

Eprint and DOI entries are still present for conference proceedings and all other article types.

It would be really nice if MN followed Elsevier and simply showed the DOI and arXiv values directly :smile:

[keithtsmith]

I checked with the publishers - the correct format for 'in press' papers which also have an arxiv number is (example is in press at MNRAS):

Authors et al., 2014, MNRAS, preprint (arxiv:0123.4567)

[timj]

ADS aren't helping here. For ADS the journal is "ArXiv e-prints" which can be mapped fairly easily to "preprint" and "(arXiv:0123.4567)" but there is nowhere for the submitted journal to show up. All ADS arXiv entries will need modification.

Of course, if we modify the journal entry to say "MNRAS" how does the bibtex style file know when the arXiv identifier should be used? Is it meant to see if there is a volume number and only print the arXiv identifier if volume is not set?

People are noticing again that the current version of mn2e.bst is completely broken for arXiv (you just get the ArXiv journal name). I just received a patch that attempts to print arXiv identifiers out if the journal is ArXiv. That seems better than not printing anything at all but is still not-compliant.

[timj]

For reference I was referring to commit 78ff94338442fc7c68878b1ec9fd2053948f0bc3

[keithtsmith]

Yeah, I don't think there's a way around that problem without the authors editing their .bib files. There's simply no way for ADS to know that a paper has been accepted but not yet published!

[timj]

Ok, so you edit your bib file. How? Maybe journal has to be "MNRAS, preprint" and the presence of "preprint" forces the e-print identifier to be used? You can't just say "MNRAS" for journal as that would hide. Alternatively if both pages and volume are missing the e-print is used instead and journal is just "MNRAS"? I imagine that's doable. @nxg ?