The tyranny of deadline extensions

At least in my world, it has become normal and expected for deadlines to be extended by around a week. The only explanation given is something like ‘numerous requests by authors’. However I get the strong impression that the paper committees always intended to extend the deadline, and built it into their only schedules. So many conferences do this now that it is expected; I suspect that if a conference didn’t, they would get very few submissions.

There are particular conference seasons, and so often deadlines fall around the same date. This uncertainty can cause a lot of scheduling problems. It can also annoy those organised folks who work to original deadlines.

Most recently, a Monday deadline extension to the following Friday wasn’t announced until the Friday before. Until it was announced, I was wondering how much time I would be able to spend with my family over that weekend. To get around this kind of thing a couple of times, I have written to paper chairs a week or so before a deadline, politely asking whether their deadline will be extended, saying I have a tricky schedule. This worked once, although the other time I didn’t get a reply (unsurprisingly, the workload of a paper chair is unenviable).

So I propose a different approach; that deadline extensions are announced alongside the original deadlines, in the original call for proposals.

Obviously this makes no sense, but we (Nick Collins, Thor Magnusson and I) are trying it anyway in our call for video submissions, and it’ll be interesting to see how well it works. By pre-announcing the extension but being vague about what it will be, hopefully people will put the original deadline in their calendars and work to that. However while doing tricky scheduling they’ll be able to keep the extension in mind and avoid unwarranted stress…


  1. First off, I agree completely that this practice is really annoying.

    I always wonder how much this is tied to the number of papers submitted by a couple days before the deadline. In my horrifically unscientific appraisal there seems to be a (probably mild) inverse correlation between the acceptance rate of a conference and the likelihood of a deadline extension. The things (in my conference spread) that almost always extend are the conferences that accept considerably more papers then they reject, whereas the conferences that tend to not extend their deadlines have quite low accept rates (ie. sigchi basically never does and has an acceptance rate of around 20%). So the cynical take is that rather than ‘requested by authors’ paper extensions are a function of trying to increase the pool of paper submissions to a conference that otherwise would have few submissions.

    Though there is the other point about building a culture of assumed extensions that feeds this problem and makes it difficult to change…

    Anyway, back to writing this paper that’s due this evening (there was a 15 hour extension, that I’m basically ignoring)…

  2. Yes true, the ‘requested by multiple authors’ line probably indicates less author demand rather than more in a lot of cases. I mean, you _never_ see deadline extensions for funding calls…

  3. I was once on a conference committee and we planned—at the beginning—our deadline and extension deadline. No one on the committee seemed to think twice about the fact of the deadline being extended. When the first deadline came around, we already had about 10 more submissions than we could accept. The extension was issued (despite my questioning whether it was necessary) and we received a further 30 submissions over the two weeks of the extension.

    Part of the problem may be that, like in advertising, conference committees can’t be unilateralists; either we all extend our deadlines, or none of us do. Personally, I would much prefer that deadlines are never extended and that no one expects them to be extended. If you’re interested in presenting a paper, you submit it by the deadline. If you don’t, you’re not.

Leave a Reply

Your email address will not be published. Required fields are marked *