This article is published in the January 2007 issue.

Inventing the Future

“Don’t worry about what anybody else is going to do … The best way to predict the future is to invent it. Really smart people with reasonable funding can do just about anything that doesn’t violate too many of Newton’s Laws!” – Alan Kay, 1971

Perhaps the primary thing that excites most of us is captured in Alan’s famous quote. That was true for me in 1963 as a graduate student taking my first AI course, and in spite of all the invention that has occurred in the past forty-three years it is still the thing that I find the most exciting.

In the context of this quote, and as I leave NSF, I invite you to review with me how well CISE is positioned to invent the future. In the space available I cannot review the entire computing innovation ecology (NSF is funding an Academy study on exactly this question; see: I can, however, review the situation at NSF, which continues to be the funding mainstay of basic computer science and engineering research.

Before proceeding, we can stipulate that significantly more funding is needed and that we need to continue to make CISE a more flexible, focused, strategically acting organization. That said, I can assert (proudly) that CISE is much better positioned to help invent the future than it was five years ago. For example:

  • The innovation capacity (funding) of CISE has doubled due to the ITR program and other budget increases.
  • Its excellent staff are motivated not only to fund the best science, but to search for new and innovative directions (as reflected in a number of new activities).
  • There is a heightened understanding that CISE exists to serve the community and must be responsive and responsible in all our actions.
  • The overall attitude at all levels of CISE ensures that we are focusing our funding on the most important challenges for the future (e.g., networking, computational discovery)—not only supporting the confirmatory and incremental work that is also important.
  • There is a willingness to support new modalities of work (e.g., larger awards, groups, consortia, center-scale activities).

Obviously, these improvements have not come about in a vacuum; in the final analysis, they are due to you, the community. You should be proud of what we have done together in the recent past, and look toward a future in which our community continues to mature and perform on the national stage.

Let me turn now to some principles that I believe are illustrated in where CISE is today. Let’s start with the ITR program that originated under Ruzena Bajcsy’s guidance in 2000 and continued giving new grants through 2004. That program doubled the CISE budget in just five years, and those funds, which all remain in the CISE budget, provide us the heft and flexibility to do the things we are now pursuing. Perhaps more importantly, the vast array of new and innovative ITR projects that you carried out, often in collaboration with colleagues from other fields, pointed the way to a number of new and exciting visions for the future (not that we didn’t already have a lot). In short, we significantly expanded the space of things that don’t seem to violate Newton’s laws, and in the process showed that significant funding could be put to good use. That provides a great base for the next level of inventing the future.

I believe two principles are revealed here:

  1. The size of the total funding in CISE matters—not only to be able to provide more to the community, but in relation to other parts of NSF and other agencies.
  2. Important new directions for fundamental investigation can be uncovered when working closely with other disciplines.

In the past five years we have started a number of other new activities at NSF, including focused work on cybertrust, information integration, computing education, broadening participation, science of design, new networking research, large-scale research infrastructure, and cooperative community action. Let me say a few words about some of these.

Our effort in cybertrust (or cybersecurity or information security) doesn’t need much comment, nor does information integration, since both extend, sharpen, and expand work that was already well underway in our community and is well understood. They do, however, illustrate the value of emphasizing work in an area for some period of time, of bringing perhaps related but dispersed lines of work together, and for having sufficient funds beyond what is needed for base level support of an ongoing disciplinary area. In the same way, our efforts in information integration built on previous work (digital libraries, databases, search techniques), and primarily pointed existing work in new directions.

Focused funding on a specific topic, perhaps for a limited period, is essential.

The Science of Design (SoD) activity illustrates a different mode of activity that we have used more than in the past—providing funding for an entirely new research activity that is intended (in this case) to address a perceived need for a new (and more fundamental) direction in research. Specifically, the SoD program was started to encourage a focus on fundamentals of software development that over time (if successful) will be able to form the basis for significantly more productive means of creating complex software. This is a mode of activity that other agencies (notably DARPA) have employed to great effect in the past, and that I believe CISE must do much more of in the future.

I think that what this illustrates is:

CISE has the responsibility to articulate and fund entirely new directions.

Our ongoing effort to broaden participation in computing (the BPC program) and our recently announced effort (CPATH) to revitalize computing education illustrate this same modality of focus on a specific issue that is deemed of high importance to the field and to the Nation. At the same time, they also illustrate a characteristic that has not been common in CISE—namely, attempting to bring about demonstrable change in the state of the world, not just in the state of knowledge. As we as a community learn from these programs and how to do them better, I fervently believe that CISE must continue to be a very active partner in efforts to move our field forward.

The GENI Program exemplifies both of these principles and employs a third. Clearly, the GENI Research Program (currently FIND and SING) is focused on new and fundamental research in networking. Further, it employs a principle that I flippantly characterized in an earlier column as going for BHAGs—big, hirsute, audacious goals—not just incremental work within a known framework. We should have vision on a grand scale and not be timid in pursuing the fulfillment of that vision (“inventing the future”).

Have bold, grand visions and don’t be timid!

Commensurate with that is the other part of the GENI Program—the GENI Facility. It is big, it is audacious, and it illustrates yet another principle that CISE has not previously employed for the benefit of the computing research community—namely, that sometimes visionary science requires major investments in research equipment (instruments). To explore properly some of the research questions we generate, we must not be timid in asking for our fair share of the funds that are devoted to such scientific instruments for other fields. The GENI Program also, by the way, emphasizes the importance of experimentation in our field—a long-utilized principle that has not been so heavily employed as our systems have grown in size.

Big visions may take big investments; we shouldn’t be timid.

The final activity I want to address is the need for community action. Individual investigations and educational activities will always be the bedrock of computer science. Yet, as we mature as a field, expand in size, and grow in importance to all of society, we need to find mechanisms to bring us together for cooperative action that will ultimately benefit our entire community and the rest of society. We created the Computing Community Consortium for just this reason. See:

We are stronger together when addressing community-wide issues.

Whether it is obtaining more funding for individual investigators, reforming our part of the educational system, obtaining funds for very large projects, making sure that damaging legislation is not passed, or addressing any of many other efforts, in the halls of government and the court of public opinion we will always be stronger as a community than as individuals. This is not to say that we must have unanimity, but there is a time for debate and a time for presenting our collective opinion. Busy decision-makers who are confronted with dozens of demands or proposals for spending resources or passing legislation look for well-articulated cases that demonstrate the support of major parts of a community. The CCC is one way in which as a community we can develop visions and support, but its success will depend on your direct and sustained involvement.

So, what does all this have to do with the future?

As I leave NSF after almost five years, I have been asking myself if my efforts (and those of my colleagues) over that time have made, or will make, a difference. One way of asking that question is to ask if NSF/CISE is doing its part to help us invent the future. This short review is a small attempt to uncover some principles and abstract from the day-to-day of the grants process. Ultimately, it is up to others to answer that question definitively.

At the same time, I believe that within the constraints under which we operate, NSF/CISE is better positioned today—and thus, so is the field—than it was five years ago to help you invent the future. I am passionately optimistic about the future of our field and our ability to invent it. At the same time, it will take determined and continued work, not just by NSF/CISE but by all the actors in our community—that means you, me, and all of our organizations—working together to realize that potential in practice.

Emphasizing what I noted above, these accomplishments (and others) could not have been achieved without the help of many, many people. I want to thank the CISE administrative staff, the program directors (many of whom have been rotators) and other scientific staff and advisors; those of you who have served on the CISE Advisory Committee; and the hundreds (thousands?) who have attended workshops, sent us thoughtful comments and suggestions, and, of course, responded so well to new programs and activities. The Computing Research Association has been, appropriately, a special partner with CISE. Finally, without the wisdom, advice, insight, good spirits, and just plain hard work of Deborah Crawford, Deputy AD/CISE, none of what I have talked about here would have come about as well as it did.

In closing, let me add a personal note. In many ways, serving as AD/CISE has been absolutely the best position I have ever had. It has given me the rare opportunity to make a major difference for our community, science, and the entire Nation. It also has been extremely invigorating intellectually, collegially, and just on a workaday basis. Perks like a trip to the South Pole, visits to a number of big science sites, attending interesting functions in official Washington, and the opportunity to play a meaningful role on the international stage only add to the personal rewards. Finally, of course, is the sense of having been able to give back at least as much as the field has given to me.

It has been an honor and a privilege to serve you while at NSF, and I look forward to continuing that service as an active member of the computing community.

From the Editor: CRA is grateful to Peter Freeman for the columns he and CISE staff members have contributed to CRN over the past five years, keeping the community informed of activities at NSF/CISE. Our good wishes go with him as he joins the Washington Advisory Group in January—JS.


Inventing the Future