Tags: * poor ARR meta/review quality * for author response * for meta/reviewer discussion * distribute submissions over year * reduce in limbo times by increasing number of conferences
"ARR has serious issues about: 1) Getting paper properly reviewed on-time 2) Emergency reviews are sometimes
poor in quality. This is very understandable, but not really acceptable. 3) There is no rebuttal period,
and meta-reviewers don't facilitate discussions among reviewers Review workload is as much as the old
system, if not more (as review 5-6 papers for a conference deadline all at once, and smaller workload
afterwards). The main issue with workload is people don't really submit to ARR if there is no actual
conference deadline. Since virtual conference is common now, what if ACL creating small 1 or 2 days
virtual-only conferences in the Spring/Fall (e.g. called EMNLP Spring or ACL Fall)? These smaller conferences
can cap the paper presented at a number, rest of the qualified paper automatically goes to the standard
conference. These smaller virtual-only editions should maintain the same quality standard and papers
appear in the same proceeding as the non-virtual conference (so people are eager to submit). "