This post gives a short description of the committee’s work procedure. The first paragraph is intended for the inexperienced.

For those of you who are not familiar with how committees work, the selection process is roughly the following. Each paper is assigned to 3 PC members, who can read it or use external reviewers or do both. Reading the papers and collecting the reviews takes up most of the committee’s time, about 5 out of 8 weeks. In the remaining 3 weeks the committee discusses the papers and the reviews, and eventually selects the papers to be accepted.

There are many papers for which the reviews give a clear indication of the right decision. The hard part of the committee’s work involves discussing disputed papers and selecting among borderline papers. I should point out that every committee produces disputable decisions, and some mistakes. You can only hope to avoid the most embarrassing mistakes.

I’ll say something about borderline papers first. I was impressed by the quality of submissions, but this meant that the committee had to make rather tough decisions. We had to reject papers that received unanimously positive reviews. That’s where the borderline ended up being (we are restricted by the capacity of the schedule). The committee identified around 50 borderline papers, and we had to choose among them 5 or 6 papers. (For those familiar with EasyChair - most papers with scores 2, 1, 1 were rejected.) Choosing those 5-6 papers out of 50 or 51 papers took up about half of the discussion time, because it was indeed a difficult choice. We felt that almost all of the borderline papers could have ended up in the program. The final choice was made, in part, by assessing the “added value” to already chosen papers. For 2 of the 6 slots we ended up voting between 2-3 alternatives for each slot (papers in the same area that were thought to be of about the same quality). Aside from these few last papers, we devoted almost no attention to balancing subareas of theory. Papers were accepted based on pure merit, as judged by experts. Despite the indifference to areas, I think the program came out rather balanced between algorithms and complexity theory, with a nice presence in specialized niche areas. This is a natural outcome of a diverse committee.

Concerning disputed papers, most of the disputes are resolved once identified. There were cases where reviewers claimed to have found substantial errors in central proofs. The authors of such papers got a chance to respond to those claims (via email). The committee is not obliged to verify correctness, but if doubts arise they must be looked into. There was one paper where the dispute among experts could not be resolved. While the evaluation of that paper reached an agreement, the experts disagreed on the potential impact and influence of the result. We ended up voting on this paper (it ended up being accepted, as a judicious gamble which most people thought was worth taking).

Thus, overall 3 papers got in by majority vote, and 86 by consensus.

My next post (tomorrow or later) will be on the evaluation of papers.

This is the first of a short series of posts on the work of ICALP 2016′s track A program committee that I chaired.

We just completed our discussions, and the decision notifications were sent. 

There were 320 submissions, and we accepted 89 papers.

I’d like to encourage all theoreticians to attend ICALP 2016 in Rome. Rome has art. Rome has history. Rome has art history. Rome has fashion. Rome has design. Rome has espresso, Rome has gelato. Rome has tomatoes. Rome has porcini mushrooms. Rome even has jazz clubs (not to mention opera). Best of all, the program committee worked very hard to produce an excuse to charge your grants for the trip!

The next post will be devoted to some insight into the committee’s work process.

The beginning of a new year is the perfect time to remind you and theoreticians worldwide, as you fight to regain your sobriety, that it is time to start preparing your ICALP track A submissions. ICALP cannot be better than its submissions, so please send us your best work. The deadline for submissions is February 17. The location of the conference is Rome. The submissions format has changed: most of the page is text, not margins!!! If you require more information, please visit http://www.easyconferences.eu/icalp2016/