#ReviewLog >= 10 hours for a robotics conf. paper. I was late, and knew I would be late even after I had read over the paper and made notes before the deadline. Did I contact the AE? Yes, yes I did. #xp
@amytabb That's an intense single paper commitment! What consumed all the time in this case?
@cjmuise it was a tough paper to understand -- some notation issues and also it was hard to evaluate b/c it wasn't presented as in one field and there were some ambiguities.
That time is sort of typical though -- print off the paper, read it multiple times, look up stuff, write it up in a way that hopefully makes sense to the author and editor, fight with the submission system.
I schedule at least 8 hours for every review. So I decline a lot of reviews, with what I have going on at the moment.
@cjmuise I try not to judge writing when I am reviewing, but if I can't follow the breadcrumbs then that hurts the paper.
On the 6 papers thing, last year CVPR said it may be 10 papers. I negotiated to 2 papers. This year I declined.
I've seen others w/similar or more time / review, and I think it is important to recognize this cost.
@amytabb Aye, absolutely! Some works (or types of research) just necessitate the time. I need to flat out decline largely theoretical journal papers since it clocks in at about a full weeks worth of work (just can't spare it).
@amytabb Our (field's) reviewing load is ~6 papers, so that'd amount to nearly 50hr :-O.
I guess we save a lot through common formalisms, so it's quicker to get to the heart of the contribution.