Bill Joy, Sun Microsystems, Inc., Discusses PITAC Report

By Jean E. Smith

Bill Joy, Founder and Vice President of Research at Sun Microsystems, Inc., was interviewed by CRN on November 17, 1998. Mr. Joy and Ken Kennedy, Rice University, co-chair the President's Information Technology Advisory Committee (PITAC). The PITAC's interim report was submitted in August 1998; the final report will be released in February 1999.

CRN: Why do you think "core" research in computer science and computer engineering is critically important to achieving the objectives of the PITAC's report?

Joy: Without long-term research we will lose the industry to countries that do long-term research and development. The United States is a leader in the Internet, for example, because we did the fundamental research to understand the protocols, and that led to the whole industry around the Internet. Without that kind of research it is possible that the Internet industry would not exist in the United States, but somewhere else. So it is important to our technological base, our defense needs, and our economy that fundamental research in hypergrowth areas like technology is done here.

CRN: How would you characterize the feedback and reactions to the interim PITAC report?

Joy: Very positive overall. But some people thought the research examples we used were too short-term so, if anything, they were pushing us more in the direction that we were inclined to go anyway. But it is hard to give examples of breakthrough ideas without doing the work that leads to those ideas — except retrospectively. It's like talking about something that is invisible. Some people said the report should talk more about fields like quantum computing. If you want to see what people contemporaneously think of as long-term basic research, you can go to the DARPA website and look at what they're saying and extrapolate from there. DARPA's research is as close to the edge as any.

CRN: How will the PITAC's final report differ from the interim report?

Joy: The fundamental conclusions of the report have not changed. The committee has increased the emphasis on human-computer interface research. Some topics have been pulled up to a higher level and some have been pushed down to make room. We have tried to keep the number of recommendations finite. From the feedback received, we have tried to emphasize what people thought was most important. And we added some numbers to back up some of the things we said in the report.

The big issue now is to translate the report's findings into action in Congress, and that is where the committee needs to expend its energy in 1999. We need to find bipartisan support. From some of the things I've read in the newspaper, I think there are already some positive responses on the House side, and we've already had that kind of support in the Senate. The outgoing Speaker of the House is a strong supporter of technology. Hopefully the incoming Speaker will be as well, and we'll be able to get some support from him or from other senior people in the House.

CRN: Research funding competes with many other important priorities for Congress. What do you think will convince Congress of the necessity of pursuing the PITAC vision?

Joy: Technology has been a very high-growth industry, and the growth has been disinflationary. Technology creates high-paying jobs and it keeps the industry in the United States. It helps to produce the kind of strong budget surpluses that we are seeing right now because these high-tech companies are profitable and they pay a lot of taxes. These are the types of industries that grow the economy.

Investments that the United States made in the Sputnik era are now paying off. We need to make similar kinds of investments so we will have that kind of money available to help pay for Social Security benefits when the baby-boomers retire. The only way we are going to solve that problem, other than by cutting benefits substantially, is to grow the economy. And the strongest opportunity we have for growth is in technology.

We have to take a long-term view and invest now to cover the benefits required 20 or 30 years from now. The amount of money needed is not enormous. The truth is that university research in computing and related technologies was always government-sponsored. Much of the research was related to fighting the Cold War and to building large computers and supercomputers for defense purposes. And it also was related to building a nuclear arsenal. Now, to verify the nuclear arsenal, we need to make the peacetime version of this earlier investment.

Although the investment requires a relatively small amount of money, I understand that it is always at the margin. But if we do a rate-of-return calculation on how much tax revenue this will bring in if we can imagine the industry, it will be an incredibly good investment compared with any investment in brick and mortar. It has a huge potential payoff. Not only that, but the investment will be funding education, one of the president's top goals. An educated workforce in technology is a strong asset for the nation both for defense and for growth. Even more than an investment in research, the PITAC's recommendations would be an investment in education.

CRN: The interim report emphasizes software research — how would you compare the federal investment in software research to that of companies like Sun and Microsoft?

Joy: Sun and Microsoft are investing in software research. Unfortunately, it is not really an investment in education. No matter how many people we employ in research and how much we do — and even if we make the research results available, which we do — it doesn't directly support the university system. Industry is still a net consumer of Ph.Ds and master's students, not a supplier. We do support some research in universities, but it cannot be on the scale required and it is not the right kind of money that is needed to really build the university system.

Universities need strong support. They need longer-term support than they have received, and they need support that is less tied to specific peer review. Many times the great ideas are not what peers would think of as great. To really see over the horizon, we need to fund more truly innovative (and a little bit off-the-map) research, similar to what was done in the 1960s with the moon shot and at Xerox PARC. That is why we need the virtual centers for "Expeditions" recommended in the PITAC's report. These are really about getting a group of people together to pretend what the future will be like and imagine what would be possible, and then trying to build the environment where those things are true.

This type of effort takes some focus, and it requires larger programs than you can support from a single NSF grant. The PITAC has encouraged DARPA to adopt a more long-term focus in creating these kinds of experiments. We are very pleased to see that DARPA has, in fact, already issued a BAA (Broad Agency Announcement) seeking proposals for this kind of "Expedition" idea.

CRN: So there has been a response to the report's recommendations already?

Joy: Yes, but that's the whole point of releasing an interim report. It is part of a process, and we're halfway there. The committee received very good feedback and we appreciate the time people spent providing it. We have been reacting to that feedback and revising the work. But more importantly, we are going to try to build a consensus that this is a good investment for our country for the next 20 years.

The report is not about the companies that exist now. It is not about the research Microsoft or Sun or IBM conduct or don't conduct. It is about the fact that 20 or 30 years ago we didn't have a Microsoft or a Sun or an Apple. The PITAC is interested in the next generation of these companies. We need new companies for new ideas, and we want the people who are going to start them and work at them to get an education so they can continue to build our economy. The report is about a commitment to education. It is not an enormous commitment, given that this offers the highest growth and the most disinflationary opportunity available. These are exactly the kinds of jobs we want to create.

CRN: What specific actions would you like the computing research community to take as a result of this report?

Joy: Not only do we need the government to help by making this investment in education and research; we also need the companies in our industry to take a longer-term view and to invest more in basic research. It is unfortunate that the basic research investment has declined. The market has become so competitive that everyone is doing only development and a little bit of advanced development. A lot of companies are not investing in basic research at all. They have an opportunity and an obligation to put something back, and to create foundations and sponsor university research.

I welcome public/private partnerships to do some of these things, and I think these can be part of new programs down the road. Look at all the benefits companies received from NSF's sponsorship and development of the Internet. It would be nice to see people giving things back. There are companies that do, and I think we should do more of it, but that will not replace the government's incredibly important role as the lead sponsor of research in universities. The corporate money is too little and, in general, too conditional.

CRN: Is there any fundamental message you would like to convey to the computing research community about the PITAC report?

Joy: We are where we are because of what was done a long time ago. We should all spend time talking to students, exciting students, and encouraging students of both sexes to get involved in computing. It would be great if Hollywood would make some movies where people interested in computing weren't represented as nerds. We need to tell people that computing is a great career, it is exciting, and they should think about doing research in universities where interesting things are happening — and then go start a company using new ideas.