Subscribe to the Free Print Edition!
Celebrating 25 Years

GCN Home > Forums > Future forum: Larry Brandt > Forum

Future forum: Larry Brandt

December 19, 2003 12:00 PM

Moderated by Thomas R. Temin, GCN Staff

Thomas R. Temin, GCN:

Welcome to today's Forum with Larry Brandt of the National Science Foundation. Larry oversees a program of grants to academic researchers looking for answers to government's computing problems, ranging from access for the handicapped to taxonomies for navigating large databases.

Larry, let's begin by having you describe the program, its scope in terms of dollars, where some of the grants are going and the kinds of things you have people working on.



Larry Brandt: Our goal is to support academic computer/information science researchers in joint projects with government agencies. A typical grant is about $2-300K/yr for 3 years. We have a portfolio of about 40 active grants, and a yearly budget of $8-9M. You can check our home page for details at http://www.digitalgovernment.org



Greenbelt, MD: As a scientist by training, do you feel any tension between the traditional openness and peer-review scientific processes and the government's concern about secrecy post-9/11? Are there areas of digital government research, such as data mining, that you feel should be kept out of the public eye?

Larry Brandt:

Government data itself at the detail level, as in the material gathered in the decennial national census, has often been kept confidential - there is a tension there between providing the data to those who make or study public policy, and keeping the data confidential. We have several projects in our research portfolio which are studying how we can make more data available, while still ensuring confidentiality.

There are research projects in this area being funding by the Dept. of Homeland Security - as far as I know, the researchers are not working on classified data, and the results of their research will be published publicly. This is as you might expect, because publication and public discussion are the lifeblood of the academic research enterprise. There are years of datamining research behind us, although the public only became aware in the past several years, so the cat is out of the bag.





 DC: Please explain bioinformatics in a couple of sentences--or is that possible?

Larry Brandt: Bioinformatics as a term brings two things to my mind. One is collaborative research between biologists and computer scientists. The second is using the way animals, plants, and cells handle information to inspire new kinds of computer technologies.



 Bethesda: Which agencies are best at digital government, in your opinion? Which are doing the worst?

Larry Brandt: Many of the most innovative ideas happen at the local government level. Agencies at that level are also closer to their customers. Unfortunately, they are also less likely to have the budget to implement much of their ideas. At the Federal level, agencies are large and so often not agile. Plus Federal agencies are subject to a great deal of scrutiny from the Executive Branch as administrations can turn over every 4 years, from Congress who like to make points with their constituents, and the press. None of them have much patience for a detailed answer to questions of failed projects, to progress is slower in government than in business.



 SMM: The Library of Congress is working on a national digital information infrastructure. Is NSF contributing to that?

Larry Brandt: Yes, we have been collaborating with LoC for a couple of years. We hope to have a joint call for research in the area of long-term archiving of digital material. Archiving government records for 50-100 years is an enormous upcoming problem that we need to get ahead of. There are no solutions at present, particularly for non-textual material.



 Stan in Savage, Md.: What technologies are emerging that could help federal agencies accomplish their missions more effectively and efficiently?

Larry Brandt:

There is a great deal to be done still in the areas of privacy, trust, identity, confidentiality, etc. The next great hurdle to conducting government business on line is for agencies to know with assurance who is on the other end of the line (to use an inappropriate metaphor for a packet-switched network). NSF just announced a $30M research program in CyberTrust. I hope we can find a way to get agencies to participate and partner with researchers on this topic - government has unique and important applications which must be included in this research.

The technologies associated with Digital Libraries are just becoming mature after 6 years of research. These technologies will become essential in helping citizens find and use government data - I encourage you to check out the Digital Libraries program at http://www.cise.nsf.gov





Van Hassel: There has been a historic failure of government to fund large scale scientific research that involves significant capital costs, the SSC being the tour de force. Has government willingness and political capital necessary evolved in a more positive direction since the SSC debacle?

Larry Brandt:

I'll give you my limited viewpoint. When things are seen as societally important, money and attention flows - I'm thinking here of the Y2K problem, which was very successfully handled (at least we're still here and doing business).

E-Government is not seen by Congress as something that needs that level of attention. The budgetary process was not ultimately able to provide much in the way of targeted funding for the Bush Administration's IT gem, the 25 e-government quicksilver projects (primarily these were intended to consolidate Federal Gov't services). At one point up to $200M in new funding was going to be available - I think eventually it was more like $5M.

On the other hand there are lots of good IT projects taking place within agency budgets. These projects are small enough so they don't get their heads lopped off in times of tight budgets.

As a final comment, my recollection is that the physics community never presented a unified face for the SSC (Superconducting Super Colllider) which made it easier to kill it off.





 GOC, College Park: Is anyone doing basic, as opposed to applied, research in voice-recognition technology as a bioidentity encryption tool? Is there any promise for this technology or is it too easy to crack?

Larry Brandt: This is not my area, but I just did a quick query on the NSF database of awards and found severl likely projects - try it yourself at http://www.nsf.gov/home/grants/grants_awards.htm and click on the Fielded Search text. Then go to the bottom of the page and type in the word "voice" without the quotes into the box that says Full Text Search, then punch the Search button.



 Tom Temin, GCN: Larry, a number of projects you've funded involve ways of browsing multiple databases with tools that, while termed "browsers," are very different from what people commonly think of today. Can you tell us about some of these data navigating and visualizing innovations, and how far they are from becoming products?

Larry Brandt:

Tom,

Browsing (or asking queries) across databases that were never intended to support that function is one definition for the well-known term "stovepipe". The solution is not to tear them down and build them up as one integrated database, because there will always be another incompatible database one would like to integrate later. So, we need generic tools that allow the stovepipes to continue their natural life and maintain their complexity and richness, by adding a layer (sometimes known as a wrapper) that allows inter-operation. Unfortunately, a completely generic, universally useful tool like that will not be available in my lifetime. We try to build smaller, less grand tools to make progress. Many of our researchers work directly with vendors such as Oracle, Informix, Microsoft and many others whose names escape me. Those connections are how the newest and most successful research tools make it into the commercial world for us all to use.

In terms of exciting visualizations, one of your other virtual speakers, Ben Schneiderman of Maryland, is associated with one of our grants on visualization of statistical data. The principle investigator for that grant is Gary Marchionini of the U. of North Carolina. You'll find a pointer to Gary's work at http://www.digitalgovernment.org . You may also want to check out the CopLink project at the U. of Arizona, under the direction of Hsinchun Chen.





 Tom Temin. GCN:

Thank you, Larry, for spending a highly informative hour with us. I encourage readers to check out the NSF digital government program and consider attending their excellent conference held each year in May.

Best wishes for the upcoming holidays, everyone.