I am writing an article in numerical mathematics.

I have been working on this for a long time and I think I cannot cut down on any part of the main article anymore. Yet, my advisor tells me to expand and clarify some points.

I think that the only way to accomplish this it to move some ancillary proofs to a supplement, and only report the results that are used in the main argument in the article. I cannot eliminate the ancillary proofs altogether because, even though they are minor modifications of existing literature, it does the reader a better service to at least sketch the necessary modifications.

This way, I estimate to have a supplement that is between 1 to 2 times the length of the article.

Is this bad? Are there alternatives that are not as drastic as not publishing or publishing two papers?

Important forgotten detail: I have a 20 page limit for the main article, coming from the target journal.

  • 3
    Have you already switched to the journal style? Your page count could be off in either direction if not. I had a ~17 page paper recently that was over 20 when switched to the journal style. Commented Mar 9 at 18:29
  • @CameronWilliams yes
    – Lilla
    Commented Mar 9 at 20:32
  • 5
    My main worry would be that the editor may consider this (perhaps with some justification) an illegitimate way of circumventing the 20-page limit. I would making an inquiry beforehand with someone from your target journal. Commented Mar 10 at 13:07
  • 5
    In my field (physics) it’s become a bit of an in joke that “letters” are more like two page abstracts for 20 pages of supplementary materials Commented Mar 10 at 19:41
  • 2
    I have read a 7 page paper on the error analysis of Gaussian elimination. It was a condensed version of a 30+ pages technical report. The report was a condensed version of a 130+ pages thesis. This was some time ago. Perhaps it is viable to put the majority of the material on arXiv before you submit your paper? Commented Mar 11 at 23:23

4 Answers 4


It's always hard to judge without knowing the details of the work, so I would recommend discussing this with your advisor. That said, my (physics) perspective is that having a supplement that's longer than the paper is perfectly fine as long as the split between main article and supplement makes sense. Yes, I've published such papers. Indeed, there are often lengthy details that only a subset of readers would be interested in, expanded expositions of ancillary content, etc. that can make sense to put in a supplement. The main question to ask yourself is whether the main text can stand on its own legs or not. As far as I can tell from your description, that does seem likely in this case. Of course, some content might make more sense in appendices than in a separate PDF.

  • 4
    If the journal supports appendices of course (re. the last sentence). Also in physics, I'd say the supplemental information being slightly longer is reasonably common, much longer not unknown particularly in experiments involving image data
    – Chris H
    Commented Mar 10 at 10:18
  • 1
    @ChrisH It also happens when there are many tables. I think there's a topological quantum chemistry paper that's 10 pages long, but has 180 pages or so of supplemental material as it covers a lot of symmetry groups.
    – Anyon
    Commented Mar 11 at 20:23

From a chemistry/chemical engineering perspective, the supplement being twice as long as the main text is not surprising or outlandish. For completeness and reproducibility, we must report even the most dry and boring parts, such as purification procedures, analysis techniques (both physical and computational), reactor design down to part sources, data that is summarized in a table but someone may still want in graphical form (like NMR results), etc. It is sufficient for the main text to gloss over many of these ("a 8 mL foo reactor" or "analyzed via H-NMR") if that step is in itself not particularly special or surprising, with the full details relegated to the supplement.

It can be polite to direct readers with a note ("See supplemental material for details.") Ultimately your editor and reviewers will have their own opinions about whether or not the main text requires those proofs, or if a reader will take your word for it.


[as a nonpublished reader of scientific articles] I'd love if all papers were this way

If the actual novel stuff was condensed down BUT all the anciliary work, the base proofs etc. are nevertheless present in the supplement (instead of me having to dig through all the cited papers and try to manually convert their work into what exactly you used).

So while I can't speak on how it's received "properly academically" I'd personally be a huge fan.


I cannot eliminate the ancillary proofs altogether because, even though they are minor modifications of existing literature, it does the reader a better service to at least sketch the necessary modifications.

There lies the conundrum: even if you move the ancillary proofs to the supplementary material (and you should do this, since they are ancillary, they belong to supplementary material...) you have to squeeze down to the bare minimum. Do not fear the scientific community will not trust you if you do not show all the passages, the "share or did not happen" mentality belongs to other activity having inifinite bandwidth to say nothing, science is the opposite, striving to using minimal bandwidth to convey dense messages.

So go ahead with your publication, but in the ancillary proofs show the path, not the full demonstration. If you are describing succintly the path, most of your readers will be able to tackle the demonstration on their own, with your minimal guidance.

Good luck!

You must log in to answer this question.

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