So I am trying to submit to a top-tier conference in AI. It is focused towards theoretical work. I have a couple of theorems which are the main results and I prove some other completely unrelated algebraic results as lemmas to prove these theorems. The proofs and lemmas are quite dense and quite long and despite my best efforts to clean the notation, they still look ugly. Furthermore, Its a two column format making the paper look really shabby.

I easily meet the page limit right now, but I am not sure the paper looks inviting to a reviewer. An option I am considering is to move the proofs to appendix and write good proof intuitions or sketches. But I am pretty sure I will be half to a full page under the conference page limit.

What would you suggest ? Considering the reviewers are not required to read the appendix ? Also considering the fact that papers under page limit are widely thought to be contributing less.

1 Answer 1


There is no reason or advantage to "filling up" the page limit. What you want is something readable with a good conceptual flow. So, the second option may be preferred.

Whichever you do, however, it will be the reviewers that judge. Either way may win or cause problems. It is harder for conferences, with fewer options for revision, but for a journal there are opportunities for revision if the reviewers object.

A second reason for the second option, assuming that your proof sketches are sufficiently convincing is that it will be a more comfortable read (quicker) for most reviewers, though some will want all the details. And if you point to the appendix, those folks can be satisfied as well. It would be different if you weren't providing those details at all.

Dense and turgid prose (or proofs) isn't fun to read.

You must log in to answer this question.

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