Larry Brandt (NSF) - The Digital Government Initiative

Please cite as

First let me set the record straight. I handled the paperwork, but several programs at the NSF supported this workshop:

• Digital Government Program
• Directorate for Geosciences
• Geography Program
• Digital Libraries Program

Just a bit about the management and policy context: the NSF works with the National Science and Technology Council and several of their committees. In my case, it is the committee on technology, and buried down in there is a group called the Federal Information Services and Applications Council. This is a meeting place for people from research agencies and from non-research, mission-type agencies - I know that these are not exclusive. We have an opportunity to work together in that forum. Again, there are plenty of research opportunities in digital government. The one I see most often is data integration. Almost every agency I talk to has stovepipes within their own walls, and would like to be able to work with datasets held by other agencies, which are also in stovepipes. We really want the involvement of the agencies. We are not looking just for a letter of support and I'll talk to you in three years kind of relationship. We're looking for the agencies to be involved in the concepts, in the writing of the proposal, and in the execution of the research, all the way through the project. We are also very happy to have users of the systems as part of the team.

I suppose the biggest challenge of the program is to convince the agencies that they really do need to have an IT research focus. Technologies are changing too quickly for them to do their normal planning process. It used to be that the planning process was that every 5 years you bought another IBM or mainframe computer. It's really not that way any more. Technologies are rolling over so quickly.

So we have a lot of work to do. NSF is funding basic research. I work under a new assistant director, Ruzena Bajcsy, and she uses the term "user-centered research". Her concept is, we're not asking the academics to develop operational software. What we are asking them to do is to collaborate with the users in the process of the work. She had this concept before she heard of the digital government program, so she and I get along very well on that. We have a lot of work to do working with the agencies. We hold workshops with the agencies to try to bridge that gap. We've done one called "Some Assembly Required: Building a Digital Government for the 21^st Century", which is pretty good reading on a plane. We did one in the area of GIS, jointly with OGC and the National Center for Supercomputing Applications. This one is the next in line, and we expect to continue to do workshops as we go along. We have emphasized planning and feasibility grants. We give people $40,000 and tell them to run with a good idea. We continue to talk to agency groups wherever we have an opportunity. I've presented to the FGDC a couple of times. The CIO Council is a very hard sell for this kind of thing. Their focus is Y2K and if you can find them talking past that, it's firewalls and security. They're not interested in things that are more than 6 months to a year away.

We have personnel exchange. For example, Valerie Gregg is on a 4-year detail from the Census Bureau to work with me to build the digital government program. It is really essential, not just because of the work she does, but because she speaks the right language for the statistical community.

So we have agencies participating in many different ways in the program: helping us get some domains off the ground, contributing funds, and proposal development. I use some people in the peer review process - I'm looking around and seeing a few faces I know, and you are going to be getting a letter from me with a few proposals in them, asking for help in reviewing. Agencies also provide data collection, people, anything you can think of. We're working most effectively so far with an existing group, the Interagency Council on Statistical Policy, that's headed by the chief statistician of the government who is an associate director at OMB. When we pitched this idea to her a couple of years ago, she and her group immediately decided to engage with us and spun off a little R&D working group, and it's been really essential in trying to get the connections going with the academics. So we have some promising possibilities with some other areas. Just a bit ago I met with Ann Frondorf of the Biological Resources Division of USGS. She is working with a group of agencies, including NSF, in the area of biodiversity, and they are very interested in partnering with us in a variety of different ways, so that looks quite promising. We also hope that we have the beginnings of a deep relationship with the Department of Justice.

We have gone through two rounds of proposal submissions. The first round we completed was received about a year ago. We funded 12 projects: 6 of them were planning grants and 6 of them research projects. We got 59 proposals in the second round which were submitted in middle July, and those we anticipate having decisions on in the springtime. So the number of proposals is good. We've had a good response from the community, and the stature of the people and the quality of the research proposed is quite high, so I'm quite pleased about that. We have a difficult review process in that we are trying not just to fund interesting quality research, but quality research that also has the potential for impacting government information services. We check for the research excellence first, and then for the remaining proposals we make a judgment on how we can strengthen federal information services. Reviewers come from many, many sectors, from all over the place. Tensions in the review process exist. You have proposals where there is great computer science and the biology is not particularly interesting, or visa versa. You have proposals that offer a short-term payoff versus a long-term. So these kinds of tensions show up in the program. And then, of course, we have to insure that what we are doing is legal, and that we are not taking a piece of NSF budget and spending on someone else's mission.

Funding level in 1999 was about $2 million from my program directly, and through a lot of work and collaboration with other agencies we were able to approximately double that amount.

A bit about the awards we made. What we are trying to do there is to generate an IT research kind of culture in agencies where it doesn't currently exist. And our hope there is that, as we go down the path with some of the proposals, the agencies will begin to pick up the cost of them as we go along. As usual with NSF, our funding does not go to agencies but rather to the academic research community in general, and the awards we have made are really quite strong in the area of GIS. Looking at the current round of proposals, something like one-third of them are focused on GIS or related to GIS in some way. And we had the same sort of thing in our initial set of planning grants - 4 of the 6 were GIS related. Many different agencies were involved in those planning grants. And the research grants have been all the way from $100,000 to $500,000 per year. One of those was GIS related, and I'll describe that in just a second.

I picked a couple of the awards to focus on. The first one is pretty strong statistical research. It's very important to agencies that collect statistics, particularly statistics that are related to individuals or businesses - what is called micro-data. They have a tension in their mission, which is that they are supposed to collect that stuff and keep it safe and private, and at the same time they are supposed to make as much of it as possible available to the research community in as many different forms as they can, and they really don't have very good algorithms for determining whether something is balanced one way or the other too much. So that's what this project is about. I was able to find about half of the funding for it from the statistical agencies and, believe it or not, from the General Services Administration who has a little pot of money for innovative IT research. The second one is a project we have funded with Hannan Samet from the University of Maryland on Spatial Spreadsheets and Browsers. This is a partnership with Doug Nebert and some Brazilian researchers, and I was able again to find good leverage in funding that.

I thought I would finish up by talking about the big initiative we have that used to be called "IT squared" or "IT research for the 21^st Century", but which has been renamed, because we were told we could get some new money if we renamed it. So we immediately renamed it, of course. The background here is that a couple of years ago there was a President's Information Technology Advisory Committee formed with high level people who came out with their final report earlier this year, but they had issued an interim report in February of last year in time for the President's 2000 budget. And unlike many of these committees, they were heard, and unlike what often happens in the President's budget, we were actually given  new money, not retargeted money, across 5 agencies; $366 million in this year's Presidential budget. NSF was supposed to be the lead. It wasn't clear what that meant, but we certainly had the biggest chunk of the money - $146 million. The President's budget didn't have a lot of detail about what we would do with that huge increase, and for us it really was huge. It would increase our budget in IT research by 50-60% in one year. So the agencies involved put together an implementation plan, which explains how they interpret this report and how they intended to respond to it programmatically. All of these are available through the Web. We still don't know what our budget is at this point. On the House side, which tends to be more stingy, they have proposed $35 million, of which only $15 would be new money. And at the same time, on the Senate side, they decided that they didn't like the term "IT squared", felt that was a bad program, but oh by the way, they were willing to give $121 million for IT research. So we have renamed the program "ITR" from "IT squared" - it was a no brainer! Part of the problem the initiative has had is that it was identified with Vice President Gore by the Republican Congress, and for a while they didn't like that, and then they decided that they could sort of trump him and they came up with a new name - something like NIDTR or something like that during the authorization process where they float great ideas. When you get to the appropriations process, the rubber meets the road and they have to actually come up with the money for this stuff. So they said, let's do it at $200 million a year and let's do it for 5 years. So that was a great statement of enthusiasm, and we liked that, and I think that actually resulted in some real money. So who knows what the number will be, but you would probably not be far wrong by splitting the difference between $35 million and $126 million and assuming that that's what the budget would be.

So, having gotten promising budget feedback from Congress, we went ahead and released a large program announcement. This is a different initiative than many others that we have had. In other words, the money is not going to be soaking down into existing programs. It is all going to be handled in a big pot. That's what this program announcement talks about - available on the NSF Website. It is intended to be responsive to the PITAC report, and to be consistent with the implementation plan. We will probably split about one-third on large centers, about one-third on multi-disciplinary groups, and another third going down to the program level. For the larger project, letters of intent and pre-proposals are required and letters of intent have to go in the middle of November.