Rotators---The Bridge Between DMS and The Math Sciences Community
July 15, 1998
Donald Lewis
A recent phone call to Donald Lewis at the National Science Foundation found him engaged in "wishful thinking for 2000." Reality supersedes fantasy almost immediately in this annual process, it turns out: When NSF division directors meet to begin talking about future budgets, the first thing on the agenda is to imagine what they would do if their funding were cut by 20%. But the next step is to build it back up, and here, Lewis points out, "You can't just say you're going to keep doing what you were doing before. . . . You need a kick; you need to show that you're opening new territory."
Where the new directions come from, of course, is the research community. As director of NSF's Division of Mathematical Sciences for the last three years, Lewis has clearly done a lot of thinking about mechanisms for getting ideas from the mathematical sciences community to the budget table. Critical to the process is one of his other major preoccupations at the moment: a staff of knowledgeable, creative, and dedicated program directors. (Although the focus of this article is on DMS, many of the ideas and comments pertain to all the federal agencies that fund research.)
The Division of Mathematical Sciences differs from other NSF divisions in the extent of its reliance on "rotators"---mathematicians who, on leave from their institutions, spend one or, preferably, two years at NSF managing research programs. Of the current 21 program directors, 13 are rotators and eight are permanent. The NSF administration would prefer to see a 40/60 ratio; Lewis believes that a move toward a 50/50 ratio is ideal for mathematics at this point.
Next year, as a result of the routine ending of some two-year rotator terms and the departure of a few long-time permanent staff, DMS will need at least one new director in every program and is looking to add to the permanent staff. Based on recent experiences, Lewis estimates that about 50 people will need to be contacted to fill each position.
Enormous Benefit to the Community
Given the extent of the recruitment (and training) burden, what's in it for DMS? Why not follow the lead of the divisions in which all or almost all staff are permanent? For Lewis, the answer is simple: The new ideas that he needs to have in his pocket when NSF funds are being allocated among disciplines come from a healthy balance between rotators and permanent staff. "Our funding isn't growing," he points out, "so by having program directors who are involved and who know where the discipline is going, NSF gets to go after the most exciting things."
(Explaining the complementary contributions of rotators and permanent staff, he points out that the permanent people, many of whom began as rotators and stayed on, provide invaluable continuity and understanding of the intricacies of NSF structure, and at the same time contribute to the scientific and intellectual mission of DMS, generating new ideas and seeing them through. The rotators can often provide input on cutting-edge developments, as they are liable to have taken part in them. "They are our ears into the community." All program officers help to define the DMS research portfolio---first by what proposals they fund, then by helping to make the case for new developments that merit budget increases, and finally by helping to frame major cross-disciplinary initiatives so that mathematics is included.)
"The thing I would like to impress on people," says NSF assistant director Robert Eisenstein, "is that having good people in Washington is of enormous benefit to the community." Eisenstein, a physicist, heads the Directorate for Mathematical and Physical Sciences, of which DMS is one of five divisions. "We prefer decisions about support for mathematics to be made by informed, knowledgeable people," he tells SIAM News; "if those people aren't here, we'll have to make the decisions ourselves."
"We do get good people" at NSF, he points out; "it just takes a lot of effort."
"I felt that it was time to do something for the community," says Carlos Berenstein of his decision to become a rotator in the analysis program in DMS. One of the things that attracted him to the position was the opportunity to work on the KDI initiative: "You have an opportunity to have some input into the future of mathematics; if you see an opportunity like that, you should grab it."
Well acquainted with NSF for many years from the outside, having received single-investigator grants in mathematics and having also participated in the group projects more typical of engineering, Berenstein believes that rotators bring a very particular advantage to NSF: "Of course you read proposals and you learn a lot about math, but a rotator can often have more empathy with the people submitting proposals." From the outside, he says, "it's always easy to get alienated if your proposal is turned down; but if you know there are people like you evaluating proposals, people who are temporary, you don't feel that you're just confronting a bureaucracy."
Berenstein thinks that if mathematicians kept in touch with the division, even after an award is made or declined, the added contact would be beneficial for both DMS and the investigator. Moreover, says Berenstein, mathematicians should consider participating more actively in multidisciplinary grants, which he believes would enhance their chances for getting support, as well as increase the societal value of their research.
Defining the Discipline
Beyond the knowledge that they're serving the community, what's in it for the mathematicians (almost always from academia) who decide to spend time at NSF? Mainly, according to Lewis, nothing less than the opportunity to help define the discipline. If people at NSF were living 20 years behind the times, he continues, "their disciplines would be shortchanged."
More concretely, Lewis views a two-year term as a program director as "the most fantastic job" for anyone thinking of becoming a department chair. "You learn who's who, and you broaden your view of mathematics." Certainly, he adds somewhat wryly, "you learn a lot about how to go after funding."
"Too often," Joe Jenkins agrees, "people in research don't take the time to stop and see what's going on in the rest of mathematics. . . . Here, you spend time looking beyond your own area." Jenkins has been at NSF since 1991, initially as a rotator in the analysis program; about a year and a half ago, he became a member of the permanent staff.
At the end of the summer, Sidney Graham will leave NSF for Central Michigan University, where he will chair the mathematics department. During his three years as a rotator at DMS, he has run various programs in the algebra/number theory program, and has participated in several larger programs. Of the approximately 125 proposals he handled during his first year, when he was responsible for number theory and combinatorics, Graham says, he felt completely comfortable evaluating 10, certainly no more than 20. With these jobs, he says, "you're going to have to stretch no matter who you are." But the experience stood him in good stead during interviews for various department chairmanships: Asked about evaluating the research of people not working in his specialty, he recalls, all he needed to do was to explain the nature of his job at NSF.
Hans Engler, a first-year rotator in the applied mathematics program, considers the career benefits of the experience unique: "Becoming a rotator at NSF has some of the aspects of a sabbatical. You're doing something different, and you're learning a lot. Your own research is cut back somewhat, but in the long run, I expect it to be even better for my career than a sabbatical."
Now ending his first year as a rotator in the analysis program, Bruce Palka is responsible for complex analysis (his own area), as well as for mathematical physics and operator theory. The exposure to areas of mathematics he "would never have known about . . . has been wonderful mathematically," he tells SIAM News. He has spent much of his time immersing himself in the new areas-learning "who's who" so that he can choose reviewers who can be counted on to provide substantive reviews and perusing old proposals, successful and unsuccessful, to familiarize himself with recent research trends in those fields.
"I have a different vision of mathematics now than when I arrived here," Palka says. "I certainly have a different view of science policy. . . . Here at NSF, you get to see how the big picture develops."
A few years at NSF, according to Robert Eisenstein, "provide a very good view of how your community is viewed in Washington, give you a chance to correct misperceptions of your community." Program directors have opportunities for enormous creativity, he continues. "You come here, you're in charge of a program, you make decisions that have a significant impact on your field. If you make good decisions, the field does well; if not, it suffers. . . . It's too bad more people don't see it that way."
What a Program Director Does
DMS receives about 2000 proposals a year and dispenses about 60% of the federal funds available for mathematical research at universities. Program directors handle an average of 100 proposals a year. Numbers of proposals would be even higher, Lewis believes, if mathematicians took declinations less personally. A mathematician will give up after submitting a proposal two times, he says, whereas researchers from other disciplines often rework and resubmit proposals many times.
In any case, the positions he's encouraging mathematicians to fill are demanding. "It's a heavy job," Lewis admits; "a sense of service to the community" is definitely required.
But "program directors don't just push proposals," says Deborah Lockhart, a permanent member of the DMS staff, currently in applied mathematics, who joined DMS in 1988 as a rotator. The pressures on NSF have increased, she points out, and program directors as a result are extensively involved in the development of new projects, whether in new multi-disciplinary areas or at the infrastucture level. "It's programs like these," she points out, "that bring in new resources."
"We look for people a number of years past tenure," she says, people who already have a somewhat broad view of mathematics."
Program directors do spend a lot of their time on the mechanics of proposal processing: choosing reviewers, getting the proposals out for review, deciding on the awards to be made, documenting the decisions, and writing declinations that, with special consideration for the mathematical temperament, are substantive, constructive, and diplomatic. Program directors, Lewis points out, can ignore a solid slate of bad reviews and fund a proposal. As long as they can defend their actions, he explains, the program directors have considerable authority and discretion.
NSF's SGER (Small Grants for Exploratory Research) program places the decision-making even more squarely with the program directors. Unlike regular proposals for work that represents natural growth, SGER proposals provide brief descriptions of small-scale, exploratory, high-risk projects. Such proposals are usually not reviewed by normal procedures; they need only convince the program director that the research is exploratory and risky but has significant potential impact (see the SGER section in the NSF Grant Proposal Guide).
Topics of SGER grants awarded by DMS include novel models of muscle mechanics (in cooperation with biology programs), new computational methods for simulating etching and deposition processes in semiconductor fabrication, and new analytical and computational methods for image processing. These grants are not made often, says DMS permanent staff member Michael Steuerwalt, but they can seed significant further work.
Much of a rotator's first year at NSF is devoted to learning the bureaucracy, and of course handling proposals-a learning process that Engler found very effective and rewarding. It's during the second year that rotators play more of a role in the development of new initiatives, which Lewis refers to as "the fun part." Citing opportunities in biology, geology, manufacturing, and quantum geometry, to name just a few, Lewis speculates about the "great new joint programs" to be developed, given the time. "We're also trying to do a lot of extending into applications," he says; "we want every program involved in applications. It is via applications that we convince the outside world of the relevance of mathematics."
At the moment, he says, "cross-cutting research is what's selling at OMB." Proud that basic mathematics was included as one of the themes in the FY 1999 budget, he points out that "one needs to keep pushing to keep funding growing for basic mathematics, and it is easier to do if you can point to breaking areas that sound exciting to nonspecialists. Building a budget is a selling job, and those who buy want to know that you are doing exciting new work. The fact that mathematics is underfunded compared with other fields has little impact on building the budget; on the other hand, the fallout from the Seiberg-Witten efforts was something we could sell."
Practical Considerations
Having established the rewards and challenges of the job, Lewis turns to practicalities. "No one takes a pay cut to come to NSF," he explains. If anything, people from certain regions find themselves with cost-of-living grants that greatly increase their take-home pay.
On the important question of time for research, he has strong opinions, based in part on personal experience. Program directors get a day a week for their research, he says. Some try to take Fridays off, and spend Friday and Saturday on their research; the more successful approach seems to be to take two months in the summer and get away. "A dedicated person," he says, "can move ahead in a serious research agenda." Most of the program directors contacted by SIAM News were somewhat less optimistic; representative was Hans Engler, who has done some research while at NSF, although mainly finishing up projects already under way, along with some exploratory work.
Lewis himself, respected for both his research (in number theory) and his accomplishments during 10 years as department chair at the University of Michigan, believes that people hesitate to spend time at NSF for the same reason they don't want to accept long terms as department chairs---the slowing down and potentially damaging interruptions of their research careers. But during a sabbatical seven years into his term as chair at Michigan, he was pleased to find that he could still prove theorems---"the thinking muscles had not atrophied." Now preparing to leave NSF for the retirement he was planning three years ago, he will finally have the opportunity to see how his research skills have weathered this further interruption.
Former SIAM president Margaret Wright, as chair of the advisory committee for the MPS directorate, has been an active participant in the search for Lewis's successor. She made at least 60 phone calls to encourage people to apply; only "a tiny number" of those she contacted actually went on to submit applications for the position, which is usually, though not necessarily, filled by a rotator.
Like Lewis, Wright realizes that people can have legitimate reasons for not being available at a particular time. Nonetheless, a "sense of responsibility to the community," whether on the part of an active researcher who becomes a program director, or a senior person who accepts the position of division director, would seem to override the objection she encountered most often: "But I'm at the peak of my research career; time out now would be too damaging."
One of the things she finds the "most upsetting" is when people complain about NSF practices (for example, the size of awards, specific funding decisions, new multidisciplinary programs), yet simultaneously take a dim view of service at NSF. "There's a disconnect," she says, "if people who believe that NSF plays an essential role in U.S. mathematical sciences research not only are unwilling to spend time there themselves, but don't even support their colleagues who do choose to work at NSF."
