[ The Types Forum (announcements only), 
     http://lists.seas.upenn.edu/mailman/listinfo/types-announce ]

1. random-choice is good

Since people seem to agree that the current selection process is
somewhat arbitrary, and too conservative, I think that the human
selection should only be used to select the "acceptable papers", and
then a random choice can be used to pick the N papers we can accomodate.

This has already been suggested here, where "acceptable" is used to
decide whether to include it in the proceedings and the random choice
decides whether you get a time slot, but if noone likes this option, we
could use the random choice directly to pick the set of papers to
include in the proceedings.  It may sound outrageous (to quote David
"imagine your student has one great result and one great POPL paper and
a coin flip means they don't get to ..."), but it's not clear it would
be worse than what we already have, and it would probably reduce the
load on PC members.

2. shortening presentations

To take an example from yet another field: in literature, conference
presentations don't come with papers at all.  You submit an abstract
(like half a page) and the rest is purely oral (even slides are
unusual).  It shows that proceedings and presentations really don't have
to go hand in hand.  So we may want to make the link between the two
a bit more loose.  In many ways, conference talks are ways to promote
the article: 25 minutes aren't enough to really show much more than the
very general idea and the kind of problems it might solve.
So shortening some of the talks even more might not be such a bad idea.
I might go even further and suggest we shorten all the talks down to
10minutes.  We could link that idea with stronger "sessions" that add
a longer discussion period shared among all the presenters, so the few
talks that need/deserve more time get it back in the form of
a discussion.  This might also make the conferences themselves more
lively (rather than only having life in the backroom talks).


Reply via email to