Tags: * if ARR continues, there should be only one conference, potentially with several instances * incentivize reviewers * favor/disfavor papers by good/bad reviewers for acceptance
"if ARR were to be maintained then the notion of conferences should be abolished, we should have only
one conference with 4 or 5 instances (eg ACL vol. 1, 2) and one Findings that can accept papers all
along the year. The ARR should provide acceptance decision for the next ACL volume and the author should
be allowed to present the paper at another instance (like paper accepted for Volume 1, hosted in the
US but the author would prefer to present in Asia or Europe for ecology's sake). Similarly, a paper
should be graded as ""accepted for the next ACL"", ""accepted for Findings"", ""please revise and resubmit
with a cover letter explaining the changes and presenting the previous reviews"" and ""please don't
resubmit, the paper needs a massive rethinking""). Of course there should be a quick response period
just to avoid ambiguities and clarifies some points to avoid an unnecessary revision cycle. Finally
there should be an incentive system to make sure reviews are (i) provided in time and (ii) of enough
quality: reviewers should get some credits points and assuming their paper reaches the 3/3/3 bar (or
some arbitrary threshold), should be warranted automatic acceptance to the next ACL volume. Let's call
it the ""your paper is on the verge, but given your cool reviews and the fact we never have to chase
your reviews, you're in"" strategy. Symetrically, authors of 2 lines reviews should get negative points
and should see their next submission negatively impacted, if not desk rejected. (this ""cheh! this will
teach you to spit on your colleagues' work"" strategy will need to be fine-tuned to account with multi-authorship
of course) "