For this second symposium, the paper submission process differed from the previous one in two ways: we requested full paper submissions instead of abstracts to allow for more thoroughly presented submissions, and we extended the review period from six to twelve weeks to avoid the frenzy we endured last time.
On the other hand, we succeeded in maintaining the large number of, often lengthy, reviews for each of the submissions. The review process again consisted of two rounds. During the first round, we received an average 4.7 reviews for each paper. The results of this first round were then used to select fifty-two papers for the next round. In this second round, we asked for two additional program committee reviews, and occasionally also for additional external reviews. In the end, papers in the second round got an average of 8.9 reviews, all of which were sent to the whole program committee before our meeting. We would like to thank all of our external reviewers for making this task possible and as painless as it was. Thanks!
The program committee meeting was held for a full day in San Francisco. By then, most program committe members had digested a large number of the reviews, and each paper to be discussed had been read by at least five of the committee members. This lead to often lively, but foremost informed, discussions. The result: 19 high quality papers were selected for publication in this proceedings, two of which clearly stood out and have therefore been selected as the best papers of the conference. The best paper awards go to:
Three of the accepted papers were co-authored by program committee members, out of eight such submissions. As is customary by now, program committee authors did not see the reviews, nor participate in the discussions for their papers. Finally, all papers were shepherded by one program committee member before final publication. The shepherding process, in general, resulted in much improved versions of all papers, including those that were strong in the beginning.
Of course, after putting together a technical program there is always the realization of things you could have done differently. We wish we had taken a stronger stand on accepting experience papers, and we wish we had accepted some of the less worked out, but more controversial idea-papers. However, we are happy with the result, and believe that the papers in this proceedings represent some of the best work being done in the Operating Systems field.
Many people were instrumental in helping us do our job. Early on, we could not have managed the incoming stream of submissions without the help of Ivy Jorgensen. XYZ helped with the software that handled submissions, and notifications to the authors. On the committee, Jay Lepreau helped out with all his experience from running the show last time, and Jim Gray was generous enough to host the committee meeting and then taking us out to sail on the San Francisco Bay. The Usenix staff, was professional and supportive throughout. Ellie Young deserves special thanks for always quickly responding to every single one of our requests; Judy DesHarnais XYZ; Toni Veglia and Zanna Knight helped with the promotion; Pennfield Jensen XYZ, and finally, Dan Klein, hung in there while we were pushing his tutorial deadlines as much as we possibly could. Thank you all!
Above all, we thank all the authors that submitted their work to continue making OSDI a world class conference! The authors of accepted papers endured the committee's shepherding, and we truly appreciate your effort in producing the papers we believed would make this such a strong program.
Finally, (from Karin) thanks to all the people at Xerox PARC, particularly in the Computer Science Lab, for your support and advice ... and yes, I still work for you! (From Willy) ...
Karin Petersen
Willy Zwaenepoel
October 1996