SIGOPS Annual Report

July 2000 - June 2001
Submitted by: William Weihl

Awards

In the past, the only awards made by SIGOPS involved the selection of distinguished papers by the program committee of the Symposium on Operating Systems Principles (SOSP). Last year, we established an endowed award in honor of Mark Weiser, who died in the Spring of 1999. The intent of the award is to recognize innovation and encourage young researchers. The award will first be presented at the next SOSP in 2001, and subsequently will be presented yearly (at the biannual SOSP, and at OSDI -- a USENIX conference cosponsored by SIGOPS -- in alternate years).

Technical Activities

The SIGOPS European Workshop went very well last fall. Planning is well under way for the next SOSP to be held this fall, and a committee has been chosen for the next European Workshop. OSDI was also held last fall and was extremely successful. The program was very strong, including a very well received keynote by Danny Lewin, CTO of Akamai. (Note: as many of you may know, Danny was killed on 9/11/01; he was a passenger on one of the planes that crashed into the World Trade Center. He will be greatly missed.)

Services for Members

SIGOPS continues to publish a quarterly newsletter, Operating Systems Review (OSR). William Waite (University of Colorado) continues as editor of our newsletter, which provides a valuable service to members in disseminating unrefereed work quickly.

Last year, at the initiative of the information director, Mike Dahlin, SIGOPS started a regular monthly mailing to members. The mailing was originally started to distribute announcements of interest to SIGOPS members. It has been expanded to include a "sites of the month" section with pointers to web sites of particular interest to the SIGOPS community. We hope that this monthly mailing will serve as a regular reminder to members of the role that SIGOPS plays in their professional lives, and that as a result retention will improve.

Challenges

There are two challenges facing SIGOPS: fragmentation of the field and declining membership.

Fragmentation seems like a natural progression of a field as it matures. SIGOPS has the advantage of breadth, which allows it to evolve and to embrace new areas (e.g., the web). At the same time, specialized groups tend to splinter off from SIGOPS. This has the potential to marginalize SIGOPS. So far this hasn’t happened in a significant way, but there are signs of it in some areas (e.g., in some areas the papers that are submitted to SOSP tend to be the ones that couldn’t get into the primary venue for that area). It may be better to let the mature areas splinter off while SIGOPS continues to move ahead with new areas, but there is the potential for a problem here that requires some attention to avoid.

Declining membership is a serious problem. The community of people who attend SOSP, the European Workshop, OSDI, and our other meetings is large and active. But many of those people are not SIGOPS members. The community we serve is much larger than our membership. By itself, this is not necessarily a problem, and in fact is probably typical of most such organizations. But the decline in our membership is an ongoing problem that must either be reversed or must be addressed with a fundamental shift in our model of how we operate and why we exist.

From June of 1996 to December of 2000, our membership declined from 3664 to 2246 – a drop of over 38% in 4.5 years. For the first 4 years, the decline was fairly linear (losing about 325 per year); it leveled out a bit in the second half of 2000, but that could be because of the timing of renewals. At the current rate, we will have no members at all in less than 8 years. I don’t have statistics going back more than 4 years, but when I looked at these statistics 2 years ago when I first became chair, the trend extended back as far as I could find data.

I think the time is fast approaching for a serious discussion of our raison d’etre. Ten or twenty years ago, people joined SIGOPS because it gave them access to our publications – both our newsletter and our conference proceedings. In the last ten years, information has been disseminated via the Internet far faster and earlier than we could do it. As a result, fewer people join SIGOPS for the publications. Some may join because they get a discount on conference registrations; some may join because they still like getting a paper copy of conference proceedings; some may join because they like supporting what we do. But joining is no longer necessary for people working in the field.

One possibility is to continue in our current mode. Membership will probably continue to decline, but financially this need not be a huge concern. (In fact, as we lose members our costs for mailing proceedings drop, so financially this probably doesn’t hurt us much.) Our conferences tend to break even or make a small amount of money. In addition, we have a large fund balance. So we can continue to fund various initiatives, provide student scholarships, and so on. If we do this and membership continues to decline, we will eventually evolve into a community service organization.

Another option is to aggressively recruit new members and to develop new offerings or enhancements to our existing offerings to increase the value of being a member. The executive committee of SIGOPS has considered a number of initiatives that might help in this area. Whether this will reverse the trend of declining membership, or merely slow it, is hard to predict.

Back to SIG Governing Board Report

Back to SIGOPS Page

SIGOPS FY'01 Annual Report posted by Rachael Barish