The White House Office of Management and Budget — the gatekeepers of agency budgets in the executive branch — and OSTP have issued guidance to federal science agencies (pdf, 360kb) directing them to make high end computing and cyberinfrastructure investments a priority in their FY 2006 budget requests, even at the expense of “lower-priority” research within the federal Networking and Information Technology Research and Development (NITRD) initiative.
While the Administration’s focus on high-end computing and cyberinfrastructure research is welcome news, the thought that they may ask agencies to cannibalize other research within the NITRD program is very worrisome.
Here’s the relevant language from the memo:

The Networking and Information Technology R&D (NITRD) program is a high Administration priority. While the importance of each of the NITRD program areas continues, high-end computing (supercomputing) and cyberinfrastructure R&D should be given higher relative priority due to the potential of each in furthering progress across a broad range of scientific and technological application areas. The recent report of the High-End Computing Revitalization Task Force (HECRTF) describes a coordinated R&D plan for core high-end computing technology, as well as multi-agency approaches for addressing high-end computing capability, capacity, and accessibility issues. Agency plans in high-end computing should be consistent with the HECRTF plan, emphasize coordination, leverage the efforts of all agencies and, where appropriate, provide explicit benefit to multiple agencies through coordinated multi-agency investments. Cyberinfrastructure R&D encompasses research on hardware and software tools that is aimed at strengthening the connections between new and existing computers (including supercomputers), databases, scientific instruments, researchers and facilities. By providing secure, reliable, distributed computing environments and tools that allow the science and engineering communities to produce, collect, store, communicate, analyze and quickly share huge amounts of information, improvements in cyberinfrastructure will accelerated discovery. Agency requests should reflect these two program priorities by reallocating funds from lower priority efforts.

Federal Computer Week has a story on the memo and what it means for NITRD. The story quotes NITRD’s David Nelson as saying other research areas under NITRD may “receive lessened funding” as a result of the guidance.

Funding cuts are not a certainty, Nelson said. Although his office’s components compete for money to some extent, “it isn’t necessarily a zero-sum game within the NITRD program,” he said. “There’s no hard and fast rule that it can’t exceed the 2005 budget.”
Offsets for new fiscal year 2006 priorities could be found from other areas of research, Nelson said. Still, with federal deficits mounting, “higher-priority things will probably come at the expense of lower-priority things,” he added.

Though Congress still hasn’t settled the FY 2005 budget, this is our first peek at how the FY 2006 budget will play out for R&D. This priority-setting will likely have a significant impact on the discipline in the years to come, so it’s crucially important that the computing community work with the agencies and the administration to ensure the right balances are struck.