PITAC's Interim Report: Expeditions Of An IT Kind

By Louise Arnheim

It was nearly two centuries ago that Thomas Jefferson asked Meriwether Lewis and William Clark to chart an overland route to the Pacific Ocean. The famed "expedition" was based on Jefferson's assumption that such a route could be charted and his shrewd calculation that America's trading position would be greatly improved if it could.

With that same adventurous spirit, a Presidential Advisory Committee envisions expeditions and similar ventures into unchartered domains of information technology (IT). Their findings are outlined in the "Interim Report to the President: New Federal Research Initiatives: Creating an Effective Management Structure" (http://www.hpcc.gov/ac/interim). The report's authors are the twenty-five members of the President's Information Technology Advisory Committee (PITAC).

While the report's primary purpose is in influence FY 2000 funding levels for federal R&D, its 1998 release is intended to "get discussion going," says PITAC co-chair Ken Kennedy.

In the September issue of Computing Research News, CRA Board Chairman Edward Lazowska urged the computing research community to rally behind the report's major ideas. In this issue, CRN takes a closer look at these ideas, and previews some of the likely hurdles they will face between now and next year's millennium budget vote.

Federal IT-R&D

PITAC's principal conclusion, that federal funding for information technology is "dangerously inadequate," suggests an economic future where the United States' "robust technological edge" is significantly diminished. In this regard, Juris Hartmanis, outgoing Associate Director of the Directorate for Computer and Information Sciences and Engineering (CISE), National Science Foundation (NSF), calls the report "a timely warning that this country is not investing enough in computer science and engineering research to maintain U.S. world leadership in information technology."

Overall, the Committee calls for $1 billion for IT over the next five years. But aside from funding levels for a few select proposals, the interim report abstains from naming specific budget figures (the Committee does, however, endorse the President's FY '99 budget for IT R&D, as well as the 21st Century Research Fund). According to Kennedy, it is likely that the final report (expected early next year) will supply such numbers.

However, increasing dollar amounts is not PITAC's only concern. The interim report also calls for fundamental changes in the modes and management of federal R&D. For example, much of the report talks about federal R&D machinery in need of realignment —from the current emphasis on applied research to a renewed appreciation for basic research; from sponsoring short-term single-investigator studies to investing in long-term, multiple-investigator studies; and from the current concentration on hardware to an intensive effort to develop software.

To maintain this balance across the wide spectrum of federal agencies engaged in IT-related R&D, PITAC says a single agency is needed to keep watch. "The only current feasible candidate" for this role, PITAC concludes, is the NSF. Accordingly, new IT funding would then be divided almost equally between NSF and other agencies. Within NSF itself, more than half of new funding would be put towards the modes cited above (the remainder would be allocated to "traditional programs within CISE —expanded as appropriate to projects of larger size and longer duration"). PITAC also recommends more IT representation on the National Science Board.

Research Priorities

The bulk of the report concentrates on four research priorities: software, scalability, high-end computing, and socio-economic and workforce impacts (notably, many of PITAC's suggestions are similar to those made at CRA's May 1997 workshop, "Research Challenges for the Next Generation Internet").

Software —Declaring software "the new physical infrastructure of the information age," PITAC recommends additional funding for software in computer science engineering and applications.

Historically, reports PITAC, federal programs have underestimated the amount of money needed for software development and testing. Today, demand for software outpaces U.S. ability to produce it and companies still depend on users to identify bugs.

And despite previous failures to establish national software libraries, PITAC believes these endeavors are so important that they should be tried again.

Scalability —Noting how the Internet's popularity and our dependence on it are increasing daily, PITAC warns that "we cannot safely extend what we currently know to more complex systems." PITAC calls for greater investment in "core software and communications technologies."

Additionally, the Committee recommends "broadening" the NGI infrastructure testbed to include information, commerce, and other services. It also urges formation of additional industry partnerships to finance projects on a scale large enough to examine design and deployment issues.

High-End Computing —The Advisory Committee prefaces recommendations in this area by briefly recounting the nation's evolution from high-performance computing to high-end computing. The Committee then calls for more research in:

• innovative computing technologies and architectures; and
• software for improving high-end computing performance.

Furthermore, the Committee encourages the attainment of petaops/petaflops-level performance more as a technology driver, not "as a goal unto itself."

Socio-Economic and Workforce Issues —Here, the Committee advocates greater research on the socio-economic impacts of technology adoption. In the Committee's judgment, research regarding the former has been merely speculative and more empirical data are needed.

With regard to workforce issues, two recommendations are particularly noteworthy for the computing research community. First, in an effort to increase IT literacy and access at all education levels, PITAC suggests expansion of government/university/industry partnerships. One such channel for expansion, it notes, could be the Experimental Program to Stimulate Competitive Research (EPSCoR).

The second recommendation arises from the Committee's conclusion that the current shortage of computer science Ph.D.s and faculty (and consequently, a substantial IT labor force) is due not only to salary considerations, but "the perception that universities are no longer the place where the most exciting work is being done." It proposes that increased funds for long-term research would attract and retain computer science graduate students.

Virtual Center Expeditions

Invoking the spirit of Lewis and Clark, the legend of H.G. Wells, and the imagination of D.H. Gelertner, PITAC invites researchers to "live in the technological future." By establishing "virtual center expeditions," the federal government would give researchers a mandate to explore the unknown. As Kennedy observes, "we have a long history of ideas (such as the World Wide Web) that weren't predicted." What the United States needs, he says, are the types of "wide-ranging explorations of the future that characterize Xerox PARC and MIT Project Mac."

As proposed by PITAC, each center could (through competitive bidding) receive as much as $40 million annually for ten years. The focus of such expeditions could be infrastructure-based (examples include distributed databases, tele-immersion) or discipline-based (bioinformatics, multiscale engineering).

Enabling Technology Centers

What's Up Ahead

Since the report's release, subpanels of PITAC members and non-members have been meeting to review comments on the interim report. Additionally, says Kennedy, Committee members have been interacting with the Office of Science and Technology Policy (OSTP) and the Office of Management and Budget (OMB) "so things are moving apace."

On Capitol Hill, the House Subcommittee on Basic Research held hearings on the report in early October. Subcommittee reaction to the report was generally favorable. In fact, several Members tried to prepare witnesses (among them, CRA Board Chairman Edward Lazowska) for the inevitable budget questions they would face in testifying before Congress next year.

Rep. Eddie Bernice Johnson (D-Texas) asked whether PITAC's goals could be met if new funds were not forthcoming. In other words, could existing R&D funds be reallocated and applied to long-term investments? Neal Lane (in his first appearance before the Subcommittee as Director of OSTP) replied that "while there's always an opportunity for reallocation," without additional revenue "it would be hard to make much progress in response to the recommendations of the Committee."

Rep. Gil Gutknecht (R-Minnesota), who favored PITAC's recommendations, cautioned that the Y2K problem might give other Members pause in appropriating new funds for IT R&D. "Here we are, a little more than a year away from the millennium," he said, and the Administration is requesting additional funds "for a problem which, frankly, did not exactly sneak up on us." Gutknecht, who also serves on the House Budget Committee, said "the computer industry has a little bit of a black eye right now." The Congressman added, "we need better answers for other folks."

In his opening statement Lazowska noted that a common PITAC theme was, essentially, "it's the software, stupid." Picking up on that paraphrase of the famous campaign slogan, Rep. Vernon Ehlers (R-Michigan) asked witnesses to explain the complexity of the software issue and how they anticipated closing the software gap.

Other concerns raised included NSF as the lead agency for coordinating related R&D, teacher training in IT, and the Department of Energy's Accelerated Strategic Computing Initiative (ASCI) program.

As this issue of CRN goes to press, PITAC is expected to hold another meeting to discuss the subpanels' findings. PITAC's final report is expected next February. Thus, by the time any budget figures are submitted to Congress, the 2000 Presidential Election will be ramping up.

CRN will continue to follow developments and report further on PITAC in future issues.