IACR Publication Reform :  Cryptology ePrint Archive Forum
Discussion related to IACR's current and future publications: conference proceedings, Journal of Cryptology, and revolution of IACR's publications.  
Goto Thread: PreviousNext
Goto: Forum ListMessage ListNew TopicSearchLog In
Posted by: nigel (IP Logged)
Date: 04 June 2013 11:15

I am putting together a more detailed proposal.

Some questions which come to mind which I would like feedback on are...

1) Should the workshops be included in this plan? I think they should otherwise we miss the main goal of getting IACR published (full) papers (fully) refereed, and published in a journal.
- To exclude workshops means someone could (in principle) submit an extended abstract to a workshop and then a "full version" to a conference.
- Including workshops would mean that the areas which are sometimes not so well represented at the main conferences (e.g. FSE/CHES like stuff) would get a look in.

But some people have suggested workshops should be treated differently. I would like to know what others feelings are; and reasons for this. There may be some other solution we have not thought of as to the workshop/conference division.

2) If we decouple acceptance to Proc. IACR from presentation at conferences/workshops; how should the programmes for conferences/workshops be determined? Any ideas here would be nice. Since we have a clean slate we could do anything.

3) How to deal with JoC? The idea is to have two journals like other learned societies. But we need to define clearly the roles of the two outlets. I have received some interesting ideas on this front, and wondered if others had some other ideas.

To answer some points others made in threads: The idea is not to reduce quality thresholds, but to maintain them in an environment where there is a larger community and more people working. This also means the "satellite" workshops (SAC, CT-RSA etc, Indocrypt) should also not be harmed.

Re: Questions
Posted by: cbw (IP Logged)
Date: 04 June 2013 11:43


Ad 1: Yes, they should. But they should not be forced to be part of the system. At the end of the day, they will have to find one solution or the other to live with Proc/IACR. Being part of the system can be one. If they should be treated differently is a different ball game. I'll reply to this separately (as to question 2).

Ad 2: I think we should have a couple of people who draft the programme. This could be substantially less than for the current workshops / conferences.

Ad 3: I think JoC should be part of Proc/IACR - or actually the other way around: JoC becomes JoC - Series A. And Proc/IACR becomes JoC - Series B. Series C should be for Open Source projects that are valuable for our membership.

Just an additional comment: I think the editors / reviewers of Proc/IACR should be independent from the programme committee of the workshops / conferences. Otherwise, this poses simply too many problems (who get's a chair - how many?).


Edited 1 time(s). Last edit at 04-Jun-2013 13:56 by cbw.

Re: Questions
Posted by: hoerder (IP Logged)
Date: 04 June 2013 15:06


sorry for the out-of-order answering. My answer for 2 is slightly long.

1) Generally I would prefer the workshops to join but they should be given a choice.

3) No clue. How do other disciplines with 2 journals distinguish between the journals?

2) Let the authors specify a preferred conference out of {C,EC,AC} and an alternative in case the paper is not rated as {C,EC,AC} quality. In addition to the reviews, every registered reviewer should be able to assign "likes" (or whatever you may call it) to papers (that haven't had a talk yet), irrespective whether they've reviewed the paper in order to provide a short term measure of the papers impact. (You can't like your own papers.)

Selection for mainline conferences: PCs can invite any paper where the author's haven't accepted a talk invitation yet up until x weeks (some global, fixed deadline) before the conference. Talk invitations from the preferred mainline conference can not be rejected by the authors, otherwise they have a week to accept/reject the invitation (without being penalised for a reject).

Selection for the workshops: PCs can invite any paper where the author's haven't accepted a talk invitation yet and that has passed the deadline for its preferred mainline conference up until y_{w} weeks (a workshop specific, fixed deadline) before the workshop. There is no restriction on acceptance/rejection for talk invitiations except that it has to happen within a week.

Open talk invitations/likes should be displayed to all PCs, registered reviewers in the system. Authors should see the likes of at least their paper and how it ranks. Optionally, the "likes" may age according to some formula. (If likes do age, I'd choose a simple aging formula such as counting only likes fresher than 1 year.) Another option is to remove papers that have been in the system for x years without getting and accepting talk invitations from the list of papers that compete for talk invitations.

Paper A wants [EC,TCC]. EC chooses it, case closed.
Paper B wants [EC,PKC]. EC does not choose it, PKC sends an invitation, authors accept, case closed.
Paper C wants [C,CHES]. EC chooses it, authors reject, C doesn't choose it, CHES chooses it, authors accept, case closed.
Paper D wants [C,CHES]. EC chooses it, authors reject, C chooses it, case closed.
Paper E wants [AC,PKC], AC doesn't choose it, PKC doesn't choose it, three years pass, topic suddenly en vogue, PKC, TCC and AC send invitations within one week, paper goes to AC by default, case closed.

Result: C/EC/AC get their choice of papers. The workshops get to choose from the remaining papers but don't risk loosing papers (where the authors accepted the talk invitation) to other venues. PC's know whether a paper has other open talk invitations so they can gauge the risk of not getting the paper when they send invites. The likes provide PCs an overview of the short term impact that a paper has but als gives them the flexibility to shape the conference/workshop program based on topics. (My base assumption is that PCs have a strong enough incentive to maintain a healthy and competitive research discipline so that massively malicious program selection does not happen. Program selection that is only a little bit malicious can be accepted as it doesn't affect paper publication, it only affects whether a paper gets boosted by a conference talk. Under that assumption, they only need an indication of what's trending/generating impact at the moment.)

Cheers, Simon

Please log in for posting a message. Only registered users may post in this forum.