CRN Interview with Juris Hartmanis

By Louise Arnheim

CRN recently interviewed Dr. Juris Hartmanis, Associate Director of the National Science Foundation's Directorate for Computer Science and Engineering (CISE).

Before his appointment as Associate Director last year, Dr. Hartmanis was Walter R. Reed Professor of Engineering at Cornell University. During his thirty years with Cornell, he helped establish the University's computer science department and became its first chair. Dr. Hartmanis also has worked in the private sector, spending seven years with General Electric's Research Laboratory, Schenectady, NY. Dr. Hartmanis has received numerous awards and recognitions for his work including computer science's highest honor, the Turing Award. He is a member of the National Academy of Engineering (NAE).

Summary: Having served as AD for a year, Dr. Hartmanis was pleased with his new role at CISE, the vanguard of computer science research. In response to CRN questions, he talked about both administrative and programming matters at CISE. For example, he noted how CISE is undergoing a reorganization that will involve the hiring of several key division and program directors. He also discussed programming initiatives (NSF-wide, and those unique to CISE) and emphasized the inter-relatedness of three initiatives in particular: VBNS, NGI, and Internet 2. Finally, Dr. Hartmanis praised the computer science community for its success in creating a "new" scientific discipline and encouraged members to take the additional step of serving in the federal government.

CRN:What are your thoughts today as you look back upon your first year at CISE?

HARTMANIS: I feel that the National Science Foundation is indeed the key foundation for science, engineering, and education. It is an institution with a noble mission; a mission to support the creation of knowledge and improve education. CISE is an important and great directorate. It is supporting and even guiding the development of a new, exciting intellectual discipline, as well as the creation of information technology.

There's no question in my mind about the impact of computer science and information technology on all aspects of our society. It is changing the basic process of doing science and engineering and most of our intellectual activities.

I believe that CISE is doing very well and I have enjoyed my first year as an AD.

CRN: Having been in Washington now also for a year, has anything surprised you about the way science policy is conducted?

HARTMANIS: I was surprised how much I am involved in NSF policy, debates, and decisions and also inter-agency activities; and many of these activities have increased over the past six months.

CRN:What is it like being a computer scientist serving in this very key and very senior position within the federal government?

HARTMANIS: Well, I believe that many more computer scientists should be serving in senior-level positions in government. Information technology is of utmost importance to all government agencies and activities and expertise and leadership in these areas is needed. For example, Anita Jones served with distinction as DDR&E-Defense Director of Research and Engineering-and did a particularly good job because she was a computer scientist.

Let me return to my comment that I find myself much more involved than I expected with NSF policy planning and decisions, as well as interagency activities and coordination. Since Anita Jones departed as DDR&E, I'm now serving as acting co-chair with Henry Kelly [of OSTP] of the National Science and Technology Council's Committee on Computing Information and Communications. This committee oversees the R&D programs of twelve federal agencies and coordinates research cooperation with US academies and industry.

CRN: What do you see as your programming priorities, or priorities in general for the next couple of years?

HARTMANIS: We have planned a reorganization of CISE and now it's time to implement it. To achieve that, we have to find two outstanding division directors and several new program directors. So, in some senses we are reorganizing and restaffing CISE. These are very important jobs and the computer science community should be much more involved in putting forth good candidates to serve at the National Science Foundation.

CRN:Is this an expansion; is it a sizing-down?

HARTMANIS: It's a reorganization. We are reducing CISE by one division, but there will be no radical changes to programs.

Programmatically, CISE has a rich set of initiatives; some of which are foundation-wide. For example, CISE will energetically participate in the foundation-wide initiative on knowledge and distributed intelligence (KDI). This initiative has three components: learning intelligent systems, knowledge networking, and new computational challenge. Clearly, CISE has a strong involvement with all of them. These initiatives provide a great opportunity for interdisciplinary work. They also allow CISE to impact programs in science and engineering and other disciplines.

We have the Digital Libraries II initiative, which is an inter-agency initiative. Again, CISE is one of the key players in this area.

We have the two infrastructure programs under PACI (Partnership for Advanced Computing Infrastructure), which will have two key centers: the NPACI in California and the NCSA in Illinois. With this new emphasis on partnerships, we are expecting real expansion of high-end computing in science and engineering.

With regard to networking, there are three component initiatives: VBNS (the very high bandwidth network system of NSF); Next Generation Internet; and Internet 2. These three initiatives promise the development of new technologies, new applications, and deployment of infrastructure, which is not yet available commercially.

CRN: Generally speaking, how do you see the current structure and CISE's role within NSF? Is there any change or adjustment you would make beyond the reorganization that's going on?

HARTMANIS: I believe that NSF overall is well-structured. I also believe that the foundation-wide initiatives - like the KDI and educational programs like IGERT-are a very positive development. They're good for science, engineering, and education and have a very, very positive impact on NSF internal planning, cooperation, and flexibility among programs.

In general, the foundation has three major themes: knowledge and distributed intelligence (KDI), life and earth environment (LEE), and education for the future (EFF). And all of these programs have inter-disciplinary components.

And, as I said earlier, CISE is being reorganized to reflect the development in the field and to achieve efficiencies in administration and operations. There will be no radical changes in programs. There will be more emphasis on experimental work and potential for support of larger, experimental projects.

CRN: In a similar vein, are you satisfied with current funding levels for CISE? For NSF?

HARTMANIS: With more funding, we clearly could do more valuable and exciting research and help alleviate the shortage of computer science professionals at all levels of education. There's no question that we have a tremendous potential for supporting really exciting and good research.

Within NSF - in the OMB-prescribed guidelines - CISE has been treated fairly. Clearly, we all hope that the Administration and Congress would see more clearly that support of research and education is the best possible investment in the future. The federal investment in computer science research and engineering has paid off very well and has greatly contributed to US leadership in information technology; leadership that is essential for national security, economic success and the creation of high quality jobs.

CRN: One initiative that is of particular interest to our readers is the current NGI initiative. What are your thoughts about the NGI?

HARTMANIS: Well, let me say again that federal investment in networking research and engineering has been a spectacular success. By any measure, this investment must be considered one of the most successful technology transfers of federally supported research to commercial applications. It's just difficult to think of anything comparable in success. And it has created an immense amount of wealth, high-quality jobs, and so on.

NGI is a very impressive initiative that involves intergovernmental agency planning and cooperation, and we sincerely hope that Congress will appropriate sufficient funds for its implementation.

The three complementary projects - NSF's VBNS (the very high bandwidth networking system), Next Generation Internet, and Internet 2 - can dramatically accelerate the development of networking technology, new applications, and deployment of infrastructure (and again, that infrastructure that would not be available commercially had the federal government not made the initial investment).

Also, the recently appointed Presidential Advisory Committee on High Performance Computing and Communications, Information Technology and Next Generation Internet enthusiastically endorsed the NGI project. So, I think that there is strong community support for it, and it would be appropriate if Congress would fund it sufficiently to get its implementation underway.

CRN:I wonder if you could also tell our readers more about the VBNS initiative.

HARTMANIS: As I mentioned earlier, these three projects [VBNS, NGI, Internet 2] are really complementary. VBNS is, as you know, NSF's high-bandwidth backbone that connects supercomputing centers and has access points for universities. We are now energetically pursuing a program to connect more universities to the backbone network. The backbone network VBNS will also serve as a backbone for Internet 2, which, as you know, is an initiative of over 100 participating universities.

And I think that Internet 2 can be viewed - in an oversimplified way - as basically the networking of campuses, and then connecting thru to the VBNS network. Clearly, NGI would help accelerate this process and would contribute to exploration of really high quality networks, to a thousand-fold improvement in capacity and exploring applications that cannot be developed with existing technologies.

CRN: Do you have any message in particular you'd like to convey to CRN's readership?

HARTMANIS: As a general comment, I'd like to say that we can all take great pride in the vitality, originality, and success of the computer science community. My message to the computer science community would be to carry on and participate in the creation of an exciting new science and explore the resulting information technologies and its applications. I think the computer science community already has done very well in this regard.

At the same time, I would urge the computer science community to get more involved in shaping the national science and engineering agenda. This community must make sure that it sends high-quality people to serve in key positions in Washington. CISE, for example, needs at this time two division directors and several program directors. These are very, very important jobs and they can be rewarding.