3

Making literature review I typically browse google Scholar, cite interesting work, download BibTex from scholar into my .bib file on overleaf - that is it.

Which is more than fine for me... until the publication time, when typically editors ask for detailed and professional bib entries (with DOI, issue, publisher, etc.) - which are typically missing from Scholar.

So what I do is I browse this works one-by-one and go into the publisher (elsevier, springer, taylor-francis - whatever) and download the full .bib file - replace it into my .bib file.

Which is doable... in like 3 hours of tedious manual work.

Any help on this - can this be automated?

Improve this question
1
  • 1
    What field are you in?
    – TimRias
    Commented Nov 29, 2021 at 12:46

2 Answers 2

Reset to default
3

Once you have the actual paper's DOI, you can use doi2bib.org to generate a bibtex entry of a paper that includes all of the relevant information. doi2bib is not flawless (because the DOI records are not), but I've found it sufficiently accurate and correct 90% of the time to be a very useful service.

Improve this answer
4
  • I was logging it to say exactly that! You can even write simple scripts to automate this work or use it from the command-line.
    – Clément
    Commented Nov 29, 2021 at 13:45
  • Yes, indeed. The output on the website is simple enough that it's easy to parse.
    – Wolfgang Bangerth
    Commented Nov 29, 2021 at 15:12
  • I’d suggest doing this over time as you write the paper. As you add references you double check the entry, and update your master database. The last thing you want to do when the paper is ‘done’ is check all the references.
    – Jon Custer
    Commented Nov 29, 2021 at 16:14
  • good idea. so just one missing step: how do I get doi's from my references (in Scholar it is often missing). And to get doi it's aobut the same time as to get .bib (you need to go to the publisher...)
    – Intelligent-Infrastructure
    Commented Nov 30, 2021 at 9:49
1

tl;dr This cannot be automated, because too many bib entries that can be found on the internet are inaccurate.

When I am reviewing papers, I always see the same mistakes* recur in bibliographies. This is because someone influential once made a mistake in generating a bib entry for the paper, and this bib file propagated through the internet, where it is downloaded time and again by people who want to have a quick solution for generating the bibliography.

Do the manual work. While doing so, keep two things in mind:

  1. the more often you do it, the quicker it will go: in the future it will surely cost you much less than 3 hours per paper.
  2. publication time is the final moment when you have influence over exactly what will be in the paper with your name on it. Instead of seeing it as tedious work, see it as your final chance to present your work in its best possible form.

*two specific highly-cited papers in my field each have an author with a double last name, and the bib files mistake the first half of the last name for an additional first name, which gets initialized. So instead of "P. Lastname Anotherlastname" this person is written as "P. L. Anotherlastname", which in turn screws up the alphabetical sorting of the entry.

Improve this answer
4
  • "This cannot be automated, because too many bib entries that can be found on the internet are inaccurate." -- This isn't wrong, but it's a question of where you go find these entries. If you go to the right source -- namely, the DOI record of a paper -- then most of the time you can actually get a reasonably good bibtex entry as well.
    – Wolfgang Bangerth
    Commented Nov 29, 2021 at 15:14
  • @WolfgangBangerth these two statements are not incompatible. Even if you go to the original source, sometimes the bibtex entry is still wrong. Even though this works most of the time, if you want to ensure a flawless bibliography in your excellent paper, you cannot get around doing some manual legwork.
    – user116675
    Commented Nov 29, 2021 at 16:32
  • Correct. But the record drawn from the DOI is often more correct than what you get doing everything by hand, or going to the authors' websites.
    – Wolfgang Bangerth
    Commented Nov 29, 2021 at 21:51
  • So, assuming both of you are right. The workflow shall be: a) while writing get .bib wherever it is easy and fast for you (e.g. Scholar), b) before publication, refine them with doi2bib - hoping to make 90% of job done and then c) finalize manually at the source.
    – Intelligent-Infrastructure
    Commented Nov 30, 2021 at 9:48

You must log in to answer this question.

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