SESSION 6 OBTAINING EXTERNAL FUNDING Susan Eggers, University of Washington Marina Chen, Yale University Helen Gigley, Naval Research Laboratory Well thanks very much to the panel. Our next session is on a topic that, as you've heard earlier today, is very important in terms of the tenure process and that is getting funding. And I think of all the things that a person has to do as an academic, this is the one that's the most like a game with strange rules, that if you know the rules you're way ahead. So, this panel is going to help you understand what some of those strange rules are. It's great to see so many of you here today. It really gives us, the presenters, a shot in the arm. We're going to try to tell you a few things about how to raise money. Our talks should all have a practical bent. The we is, I'm Susan Eggars from the University of Washington and to my right are Marina Chen from Yale University and Helen Gigley, who is now at the Naval Research Labs, but formerly worked at the National Science Foundation. And this is what we're going to talk about: we're going to cover several points in the grant submission process. (1) One of the most important is the submission strategy; basically, the networking that is necessary to get your foot in the door, what you have to do before you ever write a word. (2) Then we're going to spend some time on what a good proposal (we have no interest in bad proposals), what a good proposal looks like. (3) Then what to do when your proposal is rejected; how to turn that into a learning experience. Because, in fact, your proposals will get rejected. Ours do. All the time. Every year. We want to show you how to turn that experience around, so that you submitt a successful proposal the next time. (4) A lot of the information in these topics pertains to all funding sources: peer-reviewed federal agencies, government-reviewed federal agencies, industry, technical consortia and university grants and so forth. We also have some comments about the separate sources of funding, because they do have some different requirements. (5) And then we're going to cover many of the points I've already mentioned, but from the point of view of someone who has been on the other side, not submitting proposals, but accepting them. So, first, the submission strategy. Often people want to know whether you should submit one proposal or several. My personal opinion is the latter: that you should diversify; and there are two reasons why. The first is that not all funding sources fund everything you will need. Industry tends to give equipment, and the federal government tends to support salaries. The second reason is that you won't get all your proposals funded. So it's a good idea to cover your bets. A good model is to write what I will call a core technical proposal, a core technical text on which you can base several proposals, each of which is sent to a different source. Here are some examples. You send the very same proposal to both the National Science Foundation (whom you ask for RA salaries and your summer salary) and to Sun Microsystems (whom ask for workstations). It's the same work, but you're asking for different types of support. As another example, in the core proposal you might have different subprojects, each doing a separate part of the work, perhaps each with a different focus. You could submit some of the subprojects to the National Science Foundation and then different ones to ONR. And why do you want to do that? It saves time. If we take for given that you want to submit several proposals, it saves you time to submit variations of the same one than craft several new proposals from scratch. (Frankly, in the beginning of your career you might not have that many research ideas...!) The only time I have sent identical proposals to different federal agencies, I told them I was doing so. They said that they didn't mind. And why? Because if they both had liked the proposal, they would have funded it jointly, saving themselves some money, and getting credit for the research. Looking at it from the funders' point of view, they want to get good researchers in their programs, i.e., they want to be funding viable research. So if they can share funding with another agency, they generally like to do that. Then there's the question of whether you should be the sole principal investigator on a grant or collaborate with others; and the answer is yes. There are some agencies that have sources of funding strictly for one PI, such as the RIA, Research Initiation Award. That's the one you'll probably submit to first; it's for young faculty who have never been a PI on any other research proposal. Other NSF funds are being channeled in the direction of collaborative research, for example, the Grand Challenge funding, for which they want you to collaborate with researchers in the applications areas. So the answer is you should do both, empahsizing the approach that makes most sense for you, taking into account how far along you are in academia and how broad your research interests are. Another key piece of advise is to start the proposal-writing process early. And what do I mean by early? I mean, if these proposals are due in October or November, start gathering your thoughts in June. One reason is that it takes a long time to focus your fuzzy, new ideas, into a concrete, credible proposal. It's not merely a matter of writing things down; you're trying to motivate a problem, come up with a good solution and explain it well. It will take you several iterations to get it right. The second reason is that you're new at this. You're going to do it badly the first few times. So you want to leave yourself enough time, so that when you actually send in the proposal, it looks good. Remember, the competition for funding is high. It's important to get feedback on your proposal before you send it in. Show the proposal to someone else on your faculty, someone in your area who has written successful proposals. If there's no one on the faculty in your area, you could ask someone you met while you were interviewing, say, someone who offered you a job. You could also ask people you've met at conferences. Before you start writing, take a look at some successful proposals. Pick ones in your area, so you can at least understand what they are about, and they can serve as guides. For example, if you do experimental work, read a proposal in experimental work, so you can see what kinds of things they talk about, how much space they devote to the various sections, how they hone their arguments and so forth. I found that very helpful. Now you can begin to understand why you have to start in June. The next two points are related. The first of these is to learn which intra-organization groups fund what type of research. This applies both to subagencies or directorates in the federal government and to research groups in industry. You want to submit your proposal to the appropriate group. Meaning, you want to try to find a good match between the research you want to do and what different subgoups are interested in funding. Find that out ahead of time. (At the end of the proceedings on this session is a partial list of the federal agencies, the names of the program directors in each agency, their telephone numbers and e-mail addresses.) Should you make a mistake, particularly at NSF, your proposal will probably be routed internally to the right directorate. Try not to submit your proposal cold. Get to know your program director or industrial contact. How do you get to know him? (Pardon my sexism, but he is usually a him.) The typical way is to meet him at conferences. Many program directors -- now I'm focusing on getting federal funding -- many program directors go to conferences, primarily because they are trying to evaluate the research being presented. If you do good work, they are coming to meet you as well. You can use that opportunity to introduce yourself. Yes, I mean introduce yourself -- cold. Proposals were never granted to wallflowers. You walk right up to this guy and you say something like: "I'm Susan Eggers from the University of Washington. I'm submitting a proposal to your agency in the fall, and I'd like to talk to you about my work." or "I do work in X, Y, and Z areas; I'd like to talk to you about them, to see if you are interested in funding them." He may say, politely, and in the best case, "Oh yes, Susan. I've read your work." (...!) Don't count on that one...explain to him what you do anyway. And he may sit down with you right then and there and talk. Or he may say, "I'm too busy right now." in which case you pin him down for lunch tomorrow or after some session is over. But don't let him get away. If he just happens to be too busy at the conference, or if he's giving you the runaround, don't be flapped. Take it in stride. But get his card and tell him you'll give him a call after he gets back to Washington. And make sure you do that, and soon. No matter what he has responded, you will eventually find yourself in the situation where you are explaining your research to him. You must get across, very concisely, very coherently and very strongly what problem you are solving, your plan to solve it and why your approach is so great. You don't have much time to do this. So you should plan ahead, just like you plan for a talk. Decide what you're going to say, and, would you believe this, practice it before you pick up the telephone. Again, let me stress the main point: you have very little time to get across what your reserach is, the idea that you are competent and your research is good, and that he should fund it. Don't waste his time. And thank him when you're done. (I must sound like your mother.) As I said before, the best place to do this is at a conference, face to face. Now, why do you want to do this? There are basically two reasons. First, program directors often have funding agendas. If they like the work that you are doing, they will sometimes help you alter it a little, so that it is more in line with what they are able to fund. That's very helpful. Secondly, there are lots of proposals out there -- the funding ratio is something hideous like 1 in 5, or 1 in 8. If your proposal is one of many that have been given very good reviews, the fact that you've had a personal contact with the program director may be the tiebreaker in your favor. Don't underestimate this. Is this the most important thing I'm going to tell you? Maybe. One of them, anyway. Get to know these program directors. Often federal agencies will send out requests for proposals. They may call them something else, but that's basically what they are. Read the proposal and follow it to the letter. For example, they may ask you for a section on how your research will further undergraduate education. What on earth, you think? I devise very technical schemes for speeding up partial evaluation. Do it anyway. You will be pleasantly surprised to find that if you think about it long enough, you will discover many, many ways in which your research does, in fact, further undergraduate education. Lastly, submit on time. Often proposals are funded first come, first serve, and both federal agencies and industrial groups will run out of money. I would be a shame to be in the situation where you have a really good proposal, and there is simply no money left to fund it. So get them in on time. That's important. There are no extensions here. If you don't hear from a federal agency in a certain amount of time, say, five weeks -- you usually get a slip saying that they got your proposal and it is winding its way through the bureaucracy -- -- if you don't get one, call up the program director and ask what has gone wrong. He'll track it down. All the points on the next slide have to do with what I consider to be the heart of the proposal. This whole section contains things that are very easy for me to say but will be very hard for you to do. It's very easy for me to tell you to write a concise, original, well-argued proposal in fifteen pages, and it will be very hard for you to do that. Even the fifteen page part is hard. One important rule of thumb you can begin with is to have only one research objective for your porposal. Ask for funding for one well-defined purpose. For example, propose that you want to eliminate false sharing, not all performance issues in parallel computing, just false sharing. When you're writing about this objective, define some clearly achievable goals. You can of course put in a paragraph decribing more pie-in-the-sky goals and where they will lead to; this will show that you're thinking ahead, thinking beyond the state of the art, and that your problem and the approach that you plan to take are important enough that they might make a difference with these longer range goals. But you want to make sure that the goals you've discussed in your proposal, what you want the funding for, you can actually achieve. One of the first sections in the proposal is a discussion of the problem you plan to solve. You should make it very clear that it is a problem worth solving. The best mousetrap in the world is no good if there is no mouse. This is important even if you're working in a field where no one doubts that the problem is important. For example, some of my research involves hiding memory latencies. No one will dispute that it's an important issue, both in parallel and uniprocessor computing. I still have to say why hiding memory latencies or eliminating memory operations is appropriate in my corner of the parallel world. For example, "the performance bottleneck in low-end shared memory machines is bus traffic. Invalidation misses are responsible for most of the bus traffic; false-sharing misses comprise the largest component of that." It clearly looks like a problem. (Or at least they thought it did.) To summarize, make it very clear why the problem that you're solving is an important one to solve. Another section of the proposal should discuss how you plan to solve the problem. This discussion is fairly high level, but it still takes up most of the proposal. What you are going to do, not the nitty-gritty details of how you plan to do it. For example, "my research will identify per-process shared data and restructure that data to eliminate false sharing. We will identify unique per-process control flow graphs and extend traditional summary side-effect analysis", and then I say how I'm going to extend it. Notice I am only talking about what I'm going to do, not how I'm going to do it. I don't spend time in the proposal on details of the methodology or the experiments. And why not? Those are things that the reviewers may not understand or that may give the reviewers little points they can dispute. And, in actual fact, these are methodological details that will change over time as you gain more experience with this particular research. Why give the reviewers something they won't completely understand or can dispute, when, in fact, it's probably going to change. Now for the sell. One of the most important sections of the proposal has to sell you and your idea. This is often hard for women to do. Make sure that the following five points are part of your proposal. First, show how your work is promising. Say how right out; don't beat around the bush. Second, say how your work is different than, hopefully implying better, but maybe not, maybe just different and as also good as that of other people who are doing similar research. ("Others are transforming control in order to alter the per-process order of memory accesses. We will achieve the same ends by transforming the data directly.") An alternative would be to show how your work complements those people. Both are legitimate ways to proceed. Third, talk about how you are qualified to carry out this work. One way to do this is to discuss your track record, which, in your case, means your thesis. Another factor could be your current computing environment. "My department has a CM-5"; or "my colleagues are X and Y who are experts in Z". Any of those things can count toward how you in your current environment are qualified to carry out the work. Fourth, give the funders some indication that what you are proposing to do might actually work. If you do experimental work, put in some preliminary data. (To continue with our running proposal, "we have presented data that shows that false sharing is a performance problem, and that applying transformations manually virtually eliminates it.") If you do theory, sketch out partial proofs. If you develop algorithms, put in some algorithmic design. And last, discuss how your results will be of value to the funder. In the related work section you should show that you are up on the literature. Don't slight other researchers, but avoid gratuitous references. Yes, it's a fuzzy line. You will have probably have covered much of the related work in other sections ofthe proposal. One last word on reviewers. Sometimes you will have the opportunity to recommend reviewers for your proposal. When making suggestions, don't leave out a whole school of thought. And if you would rather the funders avoided someone, say so, and why. I have a few suggestions to keep the reviewers happy: keep to the page limits, the shorter, the better; don't give tutorials or make arguments from first principles; and add bullets to summarize the research goals near the end of the proposal. Ok, the proposal has been sent, a millenia has passed, and your reviews have now come back. Let's begin with the easy case, when it has been accepted. You should: A) Do a good job. B) Cite the funder in research papers or any other appropriate public forum. And C) Keep in touch with the program director. Sometimes that's an automatic part of your funding process; you may have to write an annual report. But if that's not a requirement, call your director; and send him your papers. It's important to keep in touch with the program director, because you would like him to fund you the next time as well. Now, when your proposal is rejected. First of all, do not be discouraged. Everyone, absolutely everyone, has proposals rejected. There is not enough money for this not to happen. So the first thing you should do is damage control on your spirits. Just take it in stride and keep submitting. Second, treat the rejection as a learning experience and use what you learn to make your next proposal that much better. One suggestion -- and this is probably only appropriate if your proposal was a borderline rejection, i.e., the reviews were pretty good reviews -- if the reviews said the proposal was a dog, don't do this -- One suggestion is to call your program director (with whom you are now of course best buddies) and ask why was it rejected. No complaining, no whining -- just a simple question. This, of course, is something you cannot do repeatedly. At NSF the program directors have to write up a summary that justifies why they have decided not to fund you. They will sometimes read that summary to you or at least use it to discuss with you what was wrong with your proposal (in their minds) and what you can do to improve it. You should always take the reviews very seriously. If they complain about something, fix it. Sometimes, the complaints result from the reviewers misunderstanding something you have written. This is probably your fault. For example, in my now famous false-sharing proposal, I was proposing to directly restructure shared data, something very different (at the time) from the rest of the community, who were indirectly restructuring data by transforming control. Every single reviewer read the proposal with control-restructuring eyes. It was simply the current strategy. Whose fault was that? Mine! So the next time I submitted that proposal, I made it very clear that I was taking an alternate tack. Let me stop here, so we can discuss differences amoung funding sources, and see how this all works from the funders point of view. What I intended to do was to give you an overview of some different funding perspectives from the government's side. If there were time, I would tell you exactly what happens to a proposal at NSF, but I don't think we'll have time for that. You've heard two different perspectives on government funding. There's basically military and non-military, and there are different ways to market your research for each. I'll cover different aspects of funding than what has been said before, and for both. First, I'll go over the agencies, but more completely. Then, where to look for the announcements. And finally, what are white papers and what are proposals? So that you get a sense of what the different jargon is, and what the different kinds of proposals are. First, the funding agencies. Some of these you've seen up here before, some you haven't, ARPA is formerly DARPA. The reason it's now called the Advance Research Project Agency is because they're now looking at dual-use research. It is no longer fashionable to be doing military research only. They have recently become the laboratory that's spearheading this. You'll find in congressional records and in the way the appropriation bills in Congress are going, everything's going dual-use. So that nothing is going to be focused on just military applications. This is your collecting up the extra military money. ONR (Office of Naval Research) is the navy research organization, and ARI is the army one. I have some more information on some of these organizations; some I don't. ARI is located in Alexandria, Virginia, if you want to try to track them down. ARPA is in Arlington, Virginia and so is ONR. They're about two blocks from each other, so that if you can come to Washington, you can visit both institutions and even NSF in a day, easily. The airforce (AFOSR) is at Boeing airforce base, which is also in D.C., but you have to have a car to get to it. The others are accessible with public transportation. Now, on to other agencies. You've heard about NSF. Thre's also NIH and NIMH. Both fund computer science research but not very much and in very constrained areas. NIMH concentrates on database research. The Department of Energy, the Department of Agriculture, just about any agency that the federal government has, funds research. So all you need to do is start calling around to find out what they're funding that you might be able to hook into. Just call. Ask. That's the best way to get a handle on it. If you have a good grants office at your university, they can put you in touch with many of these programs, and they keep in touch with special announcements that come out. There's also NASA. The Department of Commerce is another place. That's where all the translation, a large part of the natural language translation, is funded. And then, of course, NSA does a lot of the security work in coding and encryption, plus some other mathematical and theoretical kinds of computer science. Look at the list of ARPA programs at the end of the proceedings. If there's not a telephone number for your area, just call the main number. Tell them what your research area is and ask to speak to the appropriate program manager. Chances are they're on travel, because they are most of the time. If they are, see if you can get their e-mail address. The e-mail extension is still DARPA.mil; often the full address is their last name@DARPA.mil. In addition to the programs listed in the proceedings, ONR funds neuro and cognitive science, and visualization research; here they're looking at virtual environments and some other computer science-related issues. Again, call. Don't worry about calling the wrong office at NSF. Their offices are huge rooms; as soon as you call into one office, people can pass your phone call wherever it needs to go. Now, on to where you can look for funding announcements. If you want to find out what's happening in general, you need to know about the Commerce Business Daily. It comes out daily, and it's about this thick daily. It gives absolutely everything that is being advertised that you might want to consider as funding projects for any thing that's going on in the government. For the military there's something called Broad Agency Announcements (BAAs) that ONR and ARPA produce. These are basically general statements of program areas for which agencies are interested in having proposals. It's not a call for a proposal. It basically just says, "Hey, if you're doing work in this kind of an area, let us know about it. We might have some money for you." There is no commitment of funding to any of the BAAs, necessarily. So, again, you're in this selling position. I don't know how to tell you this, but as new faculty you're probably in the biggest marketing position in your life. And you might think that this is something you have no intention of doing; but let me tell you, all you're doing is marketing your ideas, yourself, your competence, your ability to get the work done. You're in a selling position from now until you get yourself established. Anyhow, the BAAs are basically obtained by schmoozing, calling up the program managers. In some cases you can't. I have some from NRL that specifically prohibit you from contacting anybody. You send in a two-page description of what you want to do. Only after someone receives the initial contact are you allowed to call to ask whether there's funding. So you have to check that on the BAAs. There are also white paper requests. (You will sometimes hear people referring to these as Requests for Proposals or RFPs.) A white paper a very general description of the research you want to do. It gives some budget information and some idea of how many students you want to support. It can't be very detailed, because it's only five pages long. Again, what it does is introduce your ideas to the program manager. This time the program manager has specific money allocated for a task and this is your way to try to get into the DARPA funding apparatus that Marina was just talking about. Often, you will find these requests for papers in the Commerce Business Daily; sometimes they come out in e-mail announcements. NSF is a very different agency. It is the only government agency that accepts unsolicited proposals. Everything that you've been hearing about is deadline-driven, meaning you have to meet the demands of somebody who's telling you "I want work done in this area". NSF, by law, cannot do this. Their charter is to process unsolicited scientific proposals. That means that you are free as scientists to send in any kind of proposal that fits within any of the auspices and programs of NSF. A program manager cannot refuse to process your proposal on technical grounds. They can refuse to process it on procedural grounds, for example, if you go over the page limit. Every proposal that NSF takes into a program must be peer reviewed. A program manager must have at least three peer reviews to be able to either refuse or actually fund a proposal. There's something called the Guide to Programs which is available from NSF. Call up the general number and ask for it. It lists all the programs at NSF; it tells you how to get on-line proposal forms; it tells you how to obtain, via e-mail, any kind of description for special initiatives, for deadline information and so on. Pick up a copy of Grants for Research in Education, Science and Engineering (GRESE). This is the NSF book that has all the forms in it for the research proposal. Now you might go to your grants office, and they'll say here are the forms. A copy of the whole booklet is what you really need, because the first half tells you what to write into the blanks on the forms, how to structure the proposal, the page limit what to emphasize and so forth. You should really read this carefully and then talk to the program director about how you might want to organize your proposal. You can use your program director at NSF to help you decide what you're going to focus on in the proposal. Another booklet that is put out by all the NSF divisions in computer science contains the abstracts of all proposals that got funding the previous year. It gives you a sense of what kinds of research NSF funds and for how much. So you might want to look at that. I'm going to touch on one other thing about writing a proposal, because I want to really make it clear. As Susan said, you need a hypothesis and a method, but you also need an evaluation metric. This is especially important, and for the entire computer scientist community. That your program works is no longer good enough to justify funding. If you think that's your evaluation metric, you'll never get anyplace. You have got to be able to demonstrate that your method worked and that it succeeded in achieving the speedup or whatever else it is that you're claiming it was supposed to do. Some final comments. On average it takes at least three resubmissions to obtain NSF funding, if you're lucky. So just know this up front. You've got at least two years probably to get your first proposal through. Unless you happen to be a special case, or unless something is really exciting about your research. That's just the way it works. The Research Initiation Award program was mentioned before. It is the one that is especially important for people who have never had federal funding. You can be tenured or non-tenured, but you cannot have had federal funding. It's meant to be startup money. Why do the program officers like it? Because it requires only 10 percent overhead to the university. Ninety percent of the funding goes to research. That's a real important bottom line. Use your grants office at your institution. Try to get them to help you find funding sources. My last point is: don't expect anybody to pat you on the head and help you out. I mean it just doesn't work that way. There's nobody who's going to say "Oh, you wrote another proposal. That's really good." So don't have this as your expectation. For some of you, graduate school has been a coddling experience. You get a lot of positive feedback, and you're really encouraged to do things. This is gone. You're on your own now. You shouldn't even complain about it. Just don't expect it to happen. And don't be shy. You really do have to take the aggressive attitude that you have got to pin people down and make sure, for instance, that you set up your appointments and do whatever else. Any questions from the floor? Hi. I've got one question for the last speaker. Is it really possible to arrange appointments with program directors? You mentioned before that people go and visit several places in Washington, D.C. during one day. Is that being done? It's done all the time; that's the expected way to work. Most universities will even fund your trips if you go to your grants office. But it's up to you to contact the program officers and set up your appointments. Basically, as a program officer at NSF, most of your work day is spent either talking on the phone or just sitting down with people and having them present what they're doing. It's very accepted. I just want to add that it's important that you get the appointment and you have prepared what you are going to say. You have about 30 seconds to get his attention, so he'll listen to you for the next half hour; they are very busy. I have a question about what happens when you get that telephone call from an NSF program director or someone else, and they say the panel has reviewed your proposal and they really like it, but they have the following concerns and would like you to cut your budget. Can you talk a little bit about that; can you give us some hints about negotiation? Answer (Eggers): I don't think it's a good idea to negotiate dollars at NSF, but you might negotiate items, which, of course, translates into dollars. Answer (Gigley): My personal experience is if they want you to cut, you take the cut. You want the grant. It's pretty much what the NSF person is telling you: "This is how many dollars I have, take it or leave it." So then you can decide what you want to change to make it fit. Now your problem in negotiation is on the university side. You need to go back to your university office and explain the situation to them and get them to kick in more, so that they're not eating it all up in overhead. I mean the real negotiation can only be at your university side. And NSF won't touch that with a ten foot pole, because they stay out of the politics at your university. You talked about proposals in general. I'm going to be writing a Research Initiation Proposal next, and I was wondering if there was anything different about those? Anything special that I should know about? I think all the topics I went through in talking about a proposal in general, the heart of the proposal, apply to RIAs. I can think of only one additional point for RIAs. They have a maximum amount of money and a maximum number of years -- go for it. But make sure that what you're proposing to do is not outrageously small for the amount of money you're asking or for the time. But this is one of the few situations other than ARPA in which you can go for the limit. One more point with this question is that the Research Initiation Award is actually evaluated by a panel instead of sending it out to arbitrary referees. It is extremely important to write a really technically sound proposal, because you have a much higher chance of getting funded through this channel. I should say that that is what we all suggest you do, apply for the RIA. It's funding geared for you; for people who have never had a grant before. So go for that first; your chances are better there. At this point my only track record is my thesis work. So do you have any comments about how much I can leverage off that and how much I should brag about that? That's your track record. Use it. It will be the same for everyone else applying for that type of grant. From eggers@cs.washington.edu Thu Jun 17 12:58:50 1993 Received: from yakima.cs.washington.edu by amber.ccs.northeastern.edu (4.1/SMI-4.1) id AA07271; Thu, 17 Jun 93 12:58:48 EDT Received: by yakima.cs.washington.edu (5.65b/7.1sun) id AA02470; Thu, 17 Jun 93 09:56:14 -0700 Date: Thu, 17 Jun 93 09:56:14 -0700 From: eggers@cs.washington.edu (Susan Eggers) Return-Path: Message-Id: <9306171656.AA02470@yakima.cs.washington.edu> To: brown@ccs.neu.edu Subject: the tables Status: R I'm sending this separately, because it is latex source. It goes at the end of the fundraising panel. Susan --------------- \documentstyle[fullpage,psfig]{article} \addtolength{\textwidth}{1in} \addtolength{\oddsidemargin}{-.4in} \addtolength{\textheight}{1.5in} \addtolength{\topmargin}{-1in} \begin{document} \pagestyle{empty} \begin{table}[here] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{National Science Foundation} \\ \multicolumn{4}{|c|}{1800 G St. NW} \\ \multicolumn{4}{|c|}{Washington, DC 20550} \\ \hline \hline \multicolumn{4}{|l|}{Directorate for Computer \& Information Science \& Engineering} \\ \hline Assistant Director&A. Nico Habermann &202-357-7936 &nhaberma@nsf.gov \\ Executive Officer&Melvin Ciment &202-357-7936 &mciment@nsf.gov \\ HPCC Coordinator&Merell Patrick &202-357-7936 &mpatrick@nsf.gov \\ Staff Associate&Jerome S. Daen &202-357-7936 &jdaen@nsf.gov \\ \hline \multicolumn{4}{|l|}{Division of Computer \& Computation Research} \\ \hline Division Director&Susan Gerhart &202-357-9747 &sgerhart@nsf.gov \\ Deputy Division Director&Bruce H. Barnes &202-357-9747 &bbarnes@nsf.gov \\ Theory of Computing PD&Dana S. Richards &202-357-7375 &richards@nsf.gov \\ Computer Systems PD&Yechezkel Zalcstein &202-357-1184 &zzalcste@nsf.gov \\ Numeric, Symbolic \& Geometric Computation PD &S. Kamal Abdali &202-357-7345 &kabdali@nsf.gov \\ Programming Languages \& Compilers PD &Forbes Lewis &202-357-7345 &flewis@nsf.gov \\ Acting Operating Systems \& Software systems PD &Forbes Lewis &202-357-7345 &flewis@nsf.gov \\ Acting Software Engineering PD&Bruce H. Barnes &202-357-9747 &bbarnes@nsf.gov \\ \hline \multicolumn{4}{|l|}{Division of Information, Robotics \& Intelligent Systems} \\ \hline Division Director&YT Chien &202-357-9572 &ytchien@nsf.gov \\ Deputy Division Director&Laurence C. Rosenberg &202-357-9592 &Irosenbe@nsf.gov \\ Database \& Expert Systems PD&Maria Zemankova &202-357-9570 &mzemanko@nsf.gov \\ Acting Information Technology \& Organizations PD & Laurence C. Rosenberg &202-357-9592 &Irosenbe@nsf.gov \\ Interactive Systems PD&Oscar Garcia &202-357-9554 &agarcia@nsf.gov \\ Knowledge Models \& Cognitive Systems PD &Su-Shing Chen &202-357-9569 &schen@nsf.gov \\ Robotics \& Machine Intelligence PD & Howard Moraff &202-357-9586 &hmoraff@nsf.gov \\ \hline \hline \multicolumn{4}{|l|}{Division of Microelectronic Information Processing Systems} \\ \hline Division Director&Bernard Chern &202-357-7373 &bchern@nsf.gov \\ Deputy Division Director&John R. Lehmann &202-357-7373 &jlehmann@nsf.gov \\ Design, Tools \& Test PD& Robert B. Grafton &202-357-7533 &rgrafton@nsf.gov \\ Acting Microelectronic Systems Architecture PD &Michael J. Foster &202-357-7853 &mfoster@nsf.gov \\ Circuits \& Signal Processing PD & John H. Cozzens &202-357-7853 &jcozzens@nsf.gov \\ Experimental Systems PD&Michael J. Foster &202-357-7853 &mfoster@nsf.gov \\ Systems Prototype \& Fabrication PD &Paul T. Hulina &202-357-7853 &phulina@nsf.gov \\ \hline \multicolumn{4}{|l|}{Division of Advanced Scientific Computing} \\ \hline Acting Division Director&Richard S. Hirsch &202-357-7558 &rhirsch@nsf.gov \\ Division Deputy Director&Vacant & & \\ Staff Associate & Stephen M. Griffin & 202-357-9776 & sgriffin@nsf.gov \\ PD, Supercomputer Centers &Michael McGrath &202-357-9776 &mmcgrath@nsf.gov \\ Associate PD, Supercomputer Centers &Lawrence E. Brandt &202-357-9776 &ilombard@nsf.gov \\ Staff Associate, Supercomputer Centers &Irene Lombardo &202-357-9776 &ilombard@nsf.gov \\ New Technologies PD&Robert G. Voigt &202-357-7727 &rvoigt@nsf.gov \\ \hline \multicolumn{4}{|l|}{Division of Networking \& Communications Research \& Infrastructure} \\ \hline Division Director&Stephen S. Wolff &202-357-9717 &steve@nsf.gov \\ Deputy Division Director&Jane C. Caviness &202-357-9717 &jcavines@nsf.gov \\ Staff Associate&Donld R. Mitchell &202-357-9717 &dmitchel@nsf.gov \\ NREN PD&Vacant & & \\ NSFnet PD&George Strawn &202-357-9717 &gstrawn@nsf.gov \\ Interagency \& International Coordinator & Steven Goldstein & 202-357-9717 &sgoldste@nsf.gov \\ Associate PD&Vacant & & \\ Associate PD&David A. Staudt &202-357-9717 &dstaudt@nsf.gov \\ Associate PD&Daniel J. Vanbelleghem &202-357-9717 &dvanbell@nsf.gov \\ Networking \& Communications Research PD &Aubrey Bush &202-357-9717 &ambush@nsf.gov \\ Associate PD&Darleen L. Fisher &202-357-9717 &dlfisher@nsf.gov \\ \hline \multicolumn{4}{|l|}{Office of Cross Disciplinary Activities} \\ \hline Head&John Cherniavsky &202-357-7349 &jchernia@nsf.gov \\ CISE Special Projects PD&Gerald L. Engel &202-357-7349 &gengel@nsf.gov \\ CISE Educational Infrastructure PD &Caroline Wardle &202-357-7349 &cwardle@nsf.gov \\ CISE Institutional Infrastructure PD &John Cherniavsky &202-357-7349 &jchernia@nsf.gov \\ CISE Cross-Directorate Activities PD &Gerald L. Engel &202-357-7349 &gengel@nsf.gov \\ CISE Research Instrumentation PD &Caroline Wardle &202-357-7349 &cwardle@nsf.gov \\ \hline \end{tabular} \end{table} \begin{table}[here] \begin{tabular}{|l|lll|} \hline %\multicolumn{4}{|c|}{Air Force Office of Scientific Research; 110 Duncan Ave., Suite B115; Bolling Air Force Base; Washington, DC 20332-0001} \\ \multicolumn{4}{|c|}{Air Force Office of Scientific Research} \\ \multicolumn{4}{|c|}{110 Duncan Ave., Suite B115} \\ \multicolumn{4}{|c|}{Bolling Air Force Base} \\ \multicolumn{4}{|c|}{Washington, DC 20332-0001} \\ \hline \hline \multicolumn{4}{|l|}{Mathematical \& Computer Sciences Directorate} \\ \hline Director & Charles J. Holland & 202-767-5025 & holland@afosr.af.mil \\ Program Manager in CS/AI & Abraham Waksman & 202-767-5028 & waksman@isi.edu \\ \hline \end{tabular} \end{table} \begin{table}[here] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{Army Research Office} \\ \multicolumn{4}{|c|}{P.O. Box 1221} \\ \multicolumn{4}{|c|}{Research Triangle Park, NC 27709-2211} \\ \hline \hline \multicolumn{4}{|l|}{Mathematical Sciences Division} \\ \hline Director, Mathematical \& Computer Science Division & Jagdish Chandra & 919-549-4254 & chandra@aro-emh1.army.mil \\ PO, Artificial Intelligence \& Software Systems & David W. Hislop & 919-549-4255 & aro@emh4.army.mil \\ PO, Numerical Analysis \& Computing & Kenneth Clark & 919-549-4256 & clark@aro-emh1.army.mil \\ \hline \end{tabular} \end{table} \begin{table}[here] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{ARPA (formerly DARPA)} \\ \multicolumn{4}{|c|}{3701 N. Fairfax Drive} \\ \multicolumn{4}{|c|}{Arlington, VA 22203} \\ \hline \hline \multicolumn{4}{|l|}{Software \& Intelligence Systems Technology Office} \\ \hline Director & Edward Thompson & 703-696-2222 & thompson@darpa.mil \\ Executive Director, Software & Vacant & & \\ Director, Computer System Technology Office & Stephen L. Squires & 703-696-2226 & squires@darpa.mil \\ Program Manager & Thomas Crystal & 703-696-2258 & crystal@darpa.mil \\ Program Manager & Paul Mockapetris & 703-696-2262 & pvm@darpa.mil \\ \hline \end{tabular} \end{table} \begin{table}[here] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{Department of Energy} \\ \multicolumn{4}{|c|}{Office of Energy Research} \\ \multicolumn{4}{|c|}{1000 Independence Ave. SW} \\ \multicolumn{4}{|c|}{Washington, DC 20585} \\ \hline \hline Director & William Happer & 202-586-5430 & not available \\ \hline \multicolumn{4}{|l|}{Office of Scientific Computing} \\ \hline Associate Director & David Nelson & 301-903-5800 & nelson@er.doe.gov \\ Deputy Associate Director & John Cavallini & 301-903-5800 & cavallini@nersc.gov \\ Program Manager & Tom Kitchens & 301-903-5800 & kitchens@er.doe.gov \\ \hline \end{tabular} \end{table} \begin{table}[here] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{NASA} \\ \multicolumn{4}{|c|}{300 E St. SW} \\ \multicolumn{4}{|c|}{Room 2R82, Code JZ} \\ \multicolumn{4}{|c|}{Washington, DC 20546} \\ \hline \hline \multicolumn{4}{|l|}{Information Systems \& Technology} \\ \hline Acting Director & Sandra Daniels-Gibson & 202-358-2215 & not available \\ Deputy Director & Vacant & & \\ \hline \multicolumn{4}{|l|}{Center of Excellence in Space Data \& Information Sciences} \\ \hline Director & Raymond Miller & 301-286-4403 & cas@cesdis1.nasa.gov \\ \hline \end{tabular} \end{table} \begin{table}[here] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{National Institute for Standards \& Technology} \\ \multicolumn{4}{|c|}{Quince Orchard and Clopper Rds.} \\ \multicolumn{4}{|c|}{Gaithersburg, MD 20899} \\ \hline \hline \multicolumn{4}{|l|}{Computer Systems Laboratory} \\ \hline Director&James H. Burrows &301-975-2822 &burrowns@micf.nist.gov \\ Associate Director&F. Lynn McNulty &301-975-3241 &mcnulty@ecf.ncsl.nist.gov \\ \hline \end{tabular} \end{table} \begin{table}[top] \begin{tabular}{|l|lll|} \hline \multicolumn{4}{|c|}{Office of Naval Research} \\ \multicolumn{4}{|c|}{800 N. Quincy St.} \\ \multicolumn{4}{|c|}{ONR Code 1133} \\ \multicolumn{4}{|c|}{Arlington, VA 22217-5660} \\ \hline \hline \multicolumn{4}{|l|}{Computer Science Division} \\ \hline Director&Andre van Tilborg &703-696-4312 &avantil@itd.nrl.navy.mil \\ Artificial Intelligence/Robotics PO &Robert Powell &703-696-4407 &powell@itd.nrl.navy.mil \\ Computer Architecture/Distributed Computing PO &Andre van Tilborg &703-696-4312 &avaintil@itd.nrl.navy.mil \\ Software Research PO &Ralph Wachter &703-696-4303 &wachter@itd.nrl.navy.mil \\ \hline %\end{tabular} %\end{table} % %\begin{table}[top] %\begin{tabular}{|l|lll|} \multicolumn{4}{c}{} \\ \multicolumn{4}{c}{} \\ \hline \multicolumn{4}{|c|}{Office of Science \& Technology Policy} \\ \multicolumn{4}{|c|}{Old Executive Office Building, Room 424} \\ \multicolumn{4}{|c|}{Washington, DC 20550;} \\ \hline \hline Assistant to the President for Science \& Technology &John H. Gibbons &202-456-7116 ¬ available \\ OSTP Director & & & \\ Associate Director&Vacant & 202-395-6175 & \\ \hline \multicolumn{4}{c}{} \\ \multicolumn{4}{c}{} \\ \multicolumn{4}{c}{} \\ \multicolumn{4}{c}{} \\ \multicolumn{4}{l}{PD = Program Director} \\ \multicolumn{4}{l}{PO = Program Officer} \end{tabular} \end{table} \end{document}