[TYPES/announce] Two phase reviewing for POPL; a response

[Alain Girault]

Jamie Andrews wrote:
> [ The Types Forum (announcements only), 
>      http://lists.seas.upenn.edu/mailman/listinfo/types-announce ]
> 
> A data point from another area that I do research in...
> 
>      The Automated Software Engineering conference (ASE) gets
> submissions in the form of 10-page papers, and accepts them as
> either "long papers" or "short papers".  Long papers get 10
> pages in the conference proceedings, and a 25-minute talk at the
> conference.  Short papers get 4 pages in the proceedings, no
> talk, and a poster at a poster session.  ASE has the poster
> session in a clean well-lighted place, in a prominent spot in
> the schedule, with no other parallel session going on, so that
> poster presenters get the most out of the discussions around
> their posters.
> 
>      I have had papers accepted there as either long or short
> papers.  I am disappointed when a paper is accepted as short,
> because it does mean it was not as well loved as the long
> papers.  On the other hand, I have had good discussions at the
> poster sessions.  A 10-minute discussion in front of a large,
> detailed poster beats a 5-minute Q&A after a talk, or a
> 20-minute discussion over lunch with only napkins to write on.
> The paper shows up on my resume as a 4-page paper at a
> conference, which is what it is, not stigmatized by a label like
> "poster".
> 
>      I'm not saying it's perfect (e.g., there is endless debate
> in the PC about what should be accepted as short and what as
> long).  However, it seems to work pretty well and gets more
> people involved than is possible when every accepted paper
> corresponds to a 25-minute talk in a session.

To avoid the problems you mention, why not allow all the papers
10 pages in the proceedings; since the proceedings are electronic,
space is not an issue; plus, reducing a paper from 10 pages to
4 pages is always very frustrating, difficult, and very long.
Then, instead of selecting oral presentation versus poster
presentation based on the ranks of the papers, do it randomly.
That way, researchers who come regularly at POPL will experience
both ways of presenting their work, and since the decision will
be made randomly, attendees will not be tempted to skip the
poster session because they will know that, statistically, 50%
of the best papers of the conference are there. Usually people
skip poster presentations because they have the wrong feeling
that these are the "low-end" papers. Random selection will
avoid that.

P.S. This 50% figure assumes that half the papers are presented
as posters, which by the way will allow you to double the number
of accepted papers.

best regards

Alain

-- 
-------------
Alain GIRAULT                       http://pop-art.inrialpes.fr/~girault
INRIA senior researcher             tel: +(33|0) 476 61 53 51
Head of the POP ART project-team    fax: +(33|0) 476 61 52 52
-----------------------------------------------------------------------------
Sauvons la Recherche ! http://www.sauvonslarecherche.fr
-----------------------------------------------------------------------------