Tags: * ACL reviewing is too complex * abolish anonymity policy * poor ARR meta/review quality
"The ACL paper submission and review process has gotten way too complicated (compared to other similar
conferences such as e.g. NeurIPS or ICML, which have to deal with many of the same problems, and at
an even larger scale), to the point where it can become a barrier of entry to the field. Please try
to simplify the process so that someone who is familiar with other CS conferences would be able to easily
navigate the ACL process, too. About opt-in vs opt-out reviewing: the most important consideration should
be the quality of the reviews. Not everyone who submits a paper to ACL is qualified to review other
submissions. I was surprised that this wasn't discussed. As an author, I would want my paper submission
to be reviewed by experts who will understand the paper, know the relevant literature, and give a thoughtful
review, even if this requires some sacrifices (such as waiting longer for reviews). I also wish there
were more discussion of the anonymity period which, while well-intentioned, just doesn't seem to work.
People are now racing to finish their papers a month before the submission deadline so they can post
their preprints to arXiv without disqualifying themselves from submitting to ACL. Therefore the policy
doesn't actually ensure anonymitySEMICOLON all it does for many authors in practice is that it shifts
the paper deadline by a month."