The CFP for this year’s Sigmetrics is now being widely circulated and it includes something very new — it takes Sigmetrics a step towards the hybrid journal/conference, a.k.a., jourference model. This represents the culmination of more than two years of discussions and work by the Sigmetrics board (of which I’m a part of), so I’m pretty excited to see how the experiment plays out!
Why go to the jourference model?
For those who have somehow managed to avoid all the debates about the pluses and minus of the conference models in CS, I won’t rehash them here. You can find in depth discussions here, here, here, and many other places…
In the case of Sigmetrics, there are additional factors at play beyond the typical ones. Sigmetrics has always been a sort of “in between” conference. It’s been a place where theoretical work, measurement work, and systems work play nicely but maintain a sort of fragile balance. This means that the PC and paper submissions are diverse, which often leads to very difficult discussions on the PC — discussions where people from different subcommunities have very different biases, styles, and opinions. As a result, there is often a situation where papers that fall in between communities (e.g. papers that mix systems and theory) get reviewers that are looking for very different things in papers, and thus have a very hard time keeping everyone happy. In a review process that doesn’t allow for revisions, this leads to a tougher path to acceptance for such papers. Yet, these are exactly the papers the community wants to support!
So, one of the key things we’re hoping to accomplish with this shift is a move from “comparing papers to one another” and looking for “reasons to reject” toward a system where reviewers ask if the paper is “interesting and high quality” and look for “reasons to accept”. Adding revisions and moving away from a model where the goal is to fill a program with 30-35 papers toward one where we accept as many papers as are interesting without trying to put them in an ordered list should lead to a situation where the program is much more diverse and provocative!
Another issue that the Sigmetrics community faces is one where many authors care about journal publications, since they overlap with the EE or OR communities. Converting Sigmetrics papers (which are longer than most journal page limits) into journal papers is difficult and frustrating…and so providing a clear, top-tier ACM journal publication for all Sigmetrics papers will hopefully help to attract more papers from these interdisciplinary researchers (a group where I include myself)!
How it will work…
To make this work for Sigmetrics we’ll be moving toward a system with three deadlines per year. Papers accepted at any point during the year will be presented at the Sigmetrics conference, and will appear in a “Proceedings of the ACM” journal immediately after acceptance (maybe before the conference). The review process for papers will be the best of journal and conference reviews. Papers will be either accepted, rejected, or offered a chance at revision. There will be only one round of revision allowed in order to ensure quick decisions on papers. Also, we will have face-to-face PC meetings (as always) in order to ensure reviewers have a chance to discuss papers in detail (and are held accountable for there interpretations of the papers).
For this year, we’ll only have time for two submission deadlines before the conference (Oct 18 and Jan 17), so be sure not to miss them! Next year will start the full three submission period cycle…
I hope to see lots of papers!