I have an article 'in press' with a journal in Elsevier. Having done some searching, it seems there are a few very similar instances/complaints that ran into the issues I am having with them. What I can't find out is: why is it so problematic (I have asked the journal manager and copyeditor, but they ignore that question).

What is the actual process used by Elsevier to take a .zip of a LaTeX project and convert it to a proof? I would think they use LaTeX, but somehow my figures get mixed up, algorithms end up with figure captions, references get removed, equations are changed, etc.

Is it someone copy-pasting and ctrl-r? There are these 'Q's that say "you didn't reference this figure", but it was referenced, with a hyperlink, the same as the other figures. So I started wondering if there really is no system, and its someone manually typing things again (which seems so unrealistic for a huge publisher).

Some of the similar questions are many years old, and it seems there is a new SkyLatex platform Elsevier mentions. I thought I would be able to make edits there, but I guess the specific journal doesn't have this option. This is also motivating my question, how or why do these multiple journals end up with different LaTeX templates than the final version and then use different backend systems?

  • 1
    I think my favourite part about Elsevier copy editing was when the journal editors forced me to re-do my submission, originally in LaTeX, in .doc format, after it already passed the first substantial part of the review (and was near accepted)... Broke my fingers getting the equations right. All equations fell apart during copy editing. I was asked to check all equations and the copy editing interface encouraged entering the corrections for complex equations as LaTeX code.
    – penelope
    Apr 12, 2021 at 13:29

1 Answer 1


Yeah the same things have happened to me a couple of times.

Most of the copy editing is outsourced and I think some employee just copy pastes your latex code into their template and does this in a very rushed and substandard manner. Occasionally they may correct a thing or two but I'm sure they do it because the typeset document looked obviously incorrect or the document did not compile with their version of tex.

The best experiences I've had with copy editing is with journals that do not use LaTeX. You are simply supposed to send your draft as a PDF and tables in MS Word. All the figures are dumped towards the end of the document and captions are placed somewhere else. Here is an example of an "Accepted Manuscript": https://joe.bioscientifica.com/view/journals/joe/aop/joe-20-0437/joe-20-0437.xml I don't even know how reviewers read this and not lose their mind when they're trying to understand Figure 3 for example. The caption is on a separate page, what the actual figure 3 is remains unknown and the location where Fig 3 is referenced in the text is 10+ pages away. So in such a scenario, the proof is actually miles better that what you had when you prepared your paper and everyone is happy :)

So yeah, going back to LaTeX typesetting, I think its just an incredibly manual process, there is no elegant "system" or automated typesetting pipeline. There are some web based typesetting clients I've seen (https://smartproof.cenveo.com/smartproof/login/sg). You see a markdown-like version of your paper but it is easier to spot mistakes, change links etc. Not sure how many people use this though.

In my view, LaTeX journal templates are just so that reviewers can read your article properly and digest it. I don't think it makes a difference when it goes to the typesetter.

You must log in to answer this question.

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