Eliminate all formatting requirements (+ survival tip)

Our conference submission was just desk rejected because the PC is unwilling to stop reading at Page 12.

We were asked to format submissions according to the LIPIcs style file, which we did, and to limit the submission length to 12 pages, excluding references; omitted proofs could be placed in the appendix. Our submission was 14.5 pages, excluding references, with no appendix.

Over the years I have submitted and also reviewed many papers that went slightly beyond the page limit; and nobody paid attention. In a few cases I have also seen PC chairs asking for the resubmission of papers which egregiously violated the formatting requirements, such as crammed, 10-point, 2-column submissions. In the present case, despite our good intentions witnessed by the usage of the LIPIcs style file, our misplacement of the \bibliography command was just too severe to be offered a second chance.

In general, there have been many discussions about formatting requirements, for example here.  The summary of my position is in the title, and details follow.

Of all the useless, time-consuming rules, the one of imposing formatting requirements strikes me as particularly outrageous because it is one that can actually be fixed. I think it is clear to everyone that a proper formatting of the paper has zero relevance compared to the myriad of actual problems that can affect submissions, such as a poorly written introduction, no intuitive exposition of the proof techniques, or missing citations. The only definite outcome of formatting requirements is that authors waste time.

I like the following paragraph from this article:

Science fiction novels of a half-century ago dramatized conflicts between humans and robots, asking if people were controlling their technologies, or if the machines were actually in charge. A few decades later, with the digital revolution in juggernaut mode, the verdict is in. The robots have won. Although the automatons were supposedly going to free people by taking on life’s menial, repetitive tasks, frequently, technological innovation actually offloads such jobs onto human beings.

I ponder upon it every time I need to waste 30 minutes with LaTeX, instead of being given the obvious option of submitting a .pdf file and possibly a .tex file from which a computer would automatically extract author names, title, and all other relevant information, including categories. They too can be quite accurately deduced from the text of the paper, can’t they? The waste of time is magnified if your paper has pictures (ours had two, carefully prepared with LaTeX extensions). No wonder we don’t see many pictures in papers.

I am not aware of any benefit of forcing papers into different latex styles, except one. Publishers can better monetize our donations if they are properly formatted. So this is one more reason to kill the current editorial system. The role of a conference/journal should be to place quality stamps on online papers, not to engulf people into a battle with latex.

Survival tip: One day I felt particularly vexed by being forced to convert all my LaTeX single-line equations, which fit nicely on a single-column format, into align environments that could fit in the 2-column proceedings format. So I offered $100 for a package that would make the process easier. I wanted a package that could recognize if I was using the “&” or the “\\” commands, and if so switch to align, or to multline. And it should also recognize if I am not using labels (in which case I need align*, not align), and so on.

Of course, a LaTeX saint quickly provided a package, and turned down my $100. I used this package ever since. I only write \[ \] and then magically the computer knows what I need. At least for this skirmish, I have won and have offloaded a mindless task onto a computer…

…though good luck making this work with the next formatting requirements.


7 thoughts on “Eliminate all formatting requirements (+ survival tip)

  1. If you’re unwilling to put in the effort to tighten the wording in your paper to get it down to the page limit, why should you expect the conference reviewers to be willing to put in the effort to read your unnecessarily bloated prose?

  2. I don’t agree with the “desk rejection”, though reviewers should not have to read arbitrarily far into a paper.

    I don’t agree that arbitrary formats are OK. There were a number of submissions to the most recent CCC that may have stated results early but really didn’t get to the point until twenty or thirty pages in; this was very frustrating and extraordinarily time-consuming as a PC member. I think that authors do owe PC members some effort in making papers accessible. It isn’t too hard to put the full paper as an appendix and then do some simple cutting away to fit the most important 12 pages or 10 pages, or whatever the reasonable reading limit is going to be. (Ideally, in the process the paper writing might improve a bit.) A latex package that made that even easier (so that “duplicate” references between the body and the appendix didn’t cause problems) would be nice, though.

    1. Thank you for sharing your perspective, Paul.

      I do not understand your experience as a CCC 2014 PC member. According to the call for papers you were not required to read past 10 pages:

      “A submission must contain within its first 10 pages a clear presentation of the merits of the paper, including discussion of its importance, prior work, and an outline of key technical ideas and methods used to achieve the main claims (similar to the content of a brief oral presentation). There is no bound on the total length of a submission, but material other than the abstract, references, and the first 10 pages is considered as supplementary and will be read at the committee’s discretion. ”

      I think these instructions are perfect.

      I find the idea that this is better accomplished by formatting requirements quite strange. (In general I cannot think of any circumstance in which it is a good idea to put the most interesting part of a paper towards the end.) And you also seem to agree that changing format is not painless.

  3. Are conferences meeting places of colleagues to exchange ideas, or are they publishers? If the former, then up to 5-7 pages of an “extended abstract” are fully enough to estimate the merits (even 2-3 would do the job, as in math). If the latter – “formatting problems” and similar are preassigned. And don’t forget the huge time-pressure on PC members (unlike for journal referees). (T)CS should just grow up from its strange treatment of the role of conferences – as publishers.

    1. Thanks Stasys for bringing up this distinction. My view of what conferences (and journals) should accomplish as publishers may be radical. As mentioned briefly in the post, I think their main function should be to place quality stamps on online papers. I plan to expand on this view in a later post.

      Two quick comments. First, the comparison with math. I heard this already, and I think it is dangerous. (Paradoxically, in my opinion) I don’t think that math is seen as such a successful field that computer science should try to imitate it.

      I also don’t see why formatting requirements are “preassigned.” Especially at the submission stage, these are in the hands of the PC. Indeed, they keep changing.

  4. O.K., let us assume that both conferences and journals are *publishing* venues. What is then the difference? (There should be some.) Hurry vs. silent decisions? Lots of “external reviewers” involved vs. 1-3? Hard deadlines vs. almost no deadlines? Announcing vs. archiving?

    Well, I see an advantage of “conference publications”: promotion of and focusing attention to “fresh” achievements. This is definitely an important (dynamic) aspect. But also a dangerous aspect: what if this (forced) focus takes a wrong route?

    P.S. I meant not ” formatting requirements”, but rather “problems like formatting”. I.e. problems naturally stipulated by the “conference hurry”. Such a process must be somehow optimized, and it is clear that the PCs are trying to do their best in this direction.

    P.P.S. On “place quality stamps on online papers”: if not just votes + or – as in, say, stackexchange – this could be an interesting alternative.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s