Newsmaker: Sun Labs' new boss

(continued from previous page)

That's a fine line you're going to be walking.
Sproull: Absolutely.

There's been a steady decline in the number of American-born Ph.D.s in computer science. Should that be a cause for concern?
Sproull: Well, yes and no.

That's a good political answer.
Sproull: I'll flesh it out a bit. From Sun's point of view I think our challenge is to really tap into talent globally and we do that in a number of ways. Some is by having engineering sites around the world and trying to hire the best engineers around the world. Another way: We get students coming from around the world, so we think of that as a summer intern program...we use community building in various ways to try to get the most passionate people working on our technology.

Where do you see the future of R&D? Is the answer to offload more and more of it to India and China?
Sproull: No, I think it has to be a mix...where the company staff is located, I think, is increasingly open.

One of the benefits of a (global) community is you don't become parochial about addressing only the needs that a local or one national group perceives. A good example is OpenOffice, which has to be an international word processor. Getting it done properly in all the different languages and documentation and so forth is done more ably by an international community.

Getting people passionate about certain kinds of device drivers or test software is really tough. And yet if you want to deliver consistent, robust performance to a customer, you can't depend on just what the community is going to be passionate about. Every community has to be supplemented by a group that makes up those gaps.

That's what part what Red Hat is doing with their distribution, they're adding the test and the quality control and filling in the gaps of the community. To some extent we do a lot of that around OpenOffice. We contribute greatly to the OpenOffice community and we do that in order to ensure the overall robustness of the result. And by the way, you can go down the major open-source things--Apache, Eclipse, NetBeans, what have you--and you'll find, in general, big companies behind most of them. I think there is a reason for that and I think that will persist. Where those company staffs are located, I think, is increasingly open. It needn't be, onshore of the domestic U.S.

But let me not skirt your question. You asked about Ph.D.s in particular. My view is that in order for the U.S. to innovate, you have to have people who are educated and trained and active in the profession. That goes back to being sure that science and technology education in primary and secondary schools is OK. In the case of Ph.D.s, it goes back to ensuring that there's academic technical research--not just in computer science research, but electrical engineering and other related things, even psychology.

President Bush came to Silicon Valley recently and talked about his American competitive initiative, which was part of the State of the Union speech. Do you think this administration really gets it or is this just more political grandstanding?
Sproull: I don't know, and I don't think I'm really qualified to answer. I think the real question is not so much whether the administration gets it as whether the Congress is beginning to get it. And I think they are. I think they are beginning to see that the pattern that gave rise to the current and immediate preceding prosperity of the U.S. in technical areas has actually been broken. And if you break it, you may not discover for 10 or 15 years that it's broken and then it's too late. I think people are beginning to understand that message. I don't think we have yet much to show for it, in terms of new research agendas and so on. I'm hopeful that the administration in talking about it will lead to something.

Let's talk about some of the toys you're working on. What's the coolest thing you've got going in the labs these days?
Sproull: Well, I don't know, if you're to measure this sort of thing by oohs and aahs, but it's got to be Sun Spot. These are small handheld battery-operated sensors with little radios on them. They're digital, they're all programmable, so it's better than Bluetooth with the complete Java environment. They're not so tiny and so starved for computational resources that you can't run Java. So this means it's very easy for people to sit at their workstation and run their NetBeans and do their Java development.

Where are you right now with turning it into a product?
Sproull: The lab is planning to sell development kits and the kit will consist, I think, of three of these Spots. You download the software onto your machine; it's essentially free. I think that the kits are $400 or $500--so just enough to cover our cost. We've got a mailing list now. We haven't actually shipped any of these yet, but very soon. We don't necessarily think that they're going to be the ultimate sensor product, but they are a way to get people interested in what it means to write Java on distributed sensor gadgets.  

More Newsmakers