Should HPCC be reauthorized?

By Fred W. Weingarten
CRA Staff

One day many years ago, when I was a young program director at the National Science Foundation, I was walking down the hall with a colleague from the social sciences division. The NSF director passed us and said, "Hi, Rick."

My friend looked at me in shock and said, "He knows you by name?" I explained that a few weeks earlier, my boss, Kent Curtis, had asked me to work with the director on some testimony regarding computers. He looked at me sympathetically and confided with the wisdom of a longtime bureaucrat, "Survival is never letting them know your name."

That attitude is not unusual in any large bureaucracy, but in this case, it was also more understandable. NSF's support of social science research funding then, as now, existed despite much criticism. Some conservative politicians believed social science research was nothing more than a pretext for expanding social programs. Others in science policy leadership positions, mainly from the physical science disciplines, saw social science as a "soft" field lacking the rigor and objectivity of "hard science," and thought it certainly unworthy of NSF funding. Given that hostile environment, my friend's perspective was rational; every time he stuck his head up, someone would shoot at it.

We in computer science, however, were trying to be noticed and taken seriously in a science community dominated by traditional physical scientists, most of whom had never seen a computer and for whom the term "computer science" seemed an oxymoron. But we were in the early stages of growth and thought we were riding the wave of the future. What we thought we needed was visibility and the chance to tell our story. Therefore, it was great news to us that the NSF director wanted to talk about computers.

I recalled that experience recently while watching the debate over whether to reauthorize the High-Performance Computing and Communications Act, which expires in the fall. The dilemma is similar. Considering all the budget-cutting knives flying on the Hill, does the community want to keep its collective head low or should it stand up and take a chance to tell the story?

To be fair, the issue is certainly more complicated, but the underlying political question of visibility lurks underneath. And the proper answer is by no means clear.

Recall, the HPC (as it was originally called--the second "C," standing for communications, was added later) was an attempt to create a coordinated, multiagency program focused on high-end computing. The issue was not so much to get agencies such as NSF or the Advanced Research Projects Agency to fund computing research--they already did. Rather, it was to establish some broader, governmentwide objectives and to direct agencies to coordinate their own research programs accordingly. By so doing, went the argument, the program would ensure more effective expenditures of computing research dollars. In addition, by linking those dollars with an explicit and politically acceptable public purpose, the program would--it was hoped--provide a basis for increased funding for computing research across the board.

Although they might quibble with details, I think most observers would judge the process a success in terms of those goals. The programs have grown rapidly within some of the participating agencies. Although it is difficult to attribute cause and effect in politics, there is little doubt in my mind that this budget growth is in great part due to public and political support of HPCC. Furthermore, coordination is much tighter among the agencies, although friction and turf battles still exist. Under the encouragement of Don Lindberg, former HPCC coordinator, and his successor, John Toole, the agencies have begun to work together more closely than they have in the past.

The basic argument against asking for reauthorization is that agency-by-agency HPCC activities now fit well within each agency's mission, and the coordination, dependent on White House leadership, has become embedded in agency culture. (That argument also was made by President Bush to a Democratic Congress, and it was not persuasive to then-Sen. Al Gore.)

The strategic reasons for deferring are twofold. First, the focus of the program has broadened immensely, to the point where it covers much of the government's investments in computing research. It has become much too large and widely framed to be encompassed under the narrow rubric of high-performance computing.

Second, there is understandable concern about opening up a congressional debate over government funding of computing research in general. Not only is Congress looking for excuses to cut the budget, but computing research is one of those fields that sits uncomfortably close to industry products and applications. It is easier to argue the importance and economic benefit from research in the area, but the argument has to be framed carefully to avoid looking like the dreaded "industrial policy."

Suppose Congress voted down a renewal bill. Would NSF, ARPA and the other agencies be expected to close down all programs in computing research--programs they now support under existing charters? It's a scary thought.

Yet it's hard to argue for standing pat on the status quo, which, quite frankly, looks dreadful for science. Under most budget proposals, overall science funding will decrease by about 30% to 40% over the next five to seven years. Some of the science agencies have been spared deep cuts this year, but that is for two reasons:

Budget cutters went after particular targets this year: the Commerce Department's Advanced Technology Program and certain environmental and energy research programs. Those easier targets will not likely be available in the future. Also, in the budget plan, the deepest cuts are saved for a few years downstream.

This line of argument says that unless the field begins to make very public and political arguments, US science will die a slow and lingering death of a thousand cuts. Computing research has a particularly good story to tell. Some form of renewal would be a good framework under which to tell that story.

The reauthorization likely would not focus on HPCC, per se, but on some broader, more general framework--for instance, information infrastructure and applications, to borrow an administration term.

It is a dilemma, and our friends in the White House and Congress are wrestling with it now. Do we hunker down and hope the winds shift before severe damage is done, with all the risks that entails? Or do we stick our heads up and engage in a fight that could well be lost? Does it even make sense to try to "sell" computing research funding on such programmatic concepts? Or is it now an established discipline that should be viewed simply as part of the nation's portfolio? We don't have much time before we'll have to decide on our answers.