Share on Facebook Tweet on Twitter Share on LinkedIn Share by email
Talking Technology with Andrew Herbert
May 13, 2010 3:22 PM PT

On New Year’s Eve, 2009, the British Queen announced her annual New Year Honours List, including recipients from all walks of life who have made a difference to their communities. Among the 2010 recipients was Andrew Herbert, Microsoft distinguished engineer and managing director of Microsoft Research Cambridge. Herbert was appointed an Officer of the Order of the British Empire (OBE) for his services to computer science.

Little more than four months later, Herbert, 56 and a native of Kent, just southeast of London, delivered the keynote speech for the second annual Enabling Innovation Through Research event, held May 5 in the Roger Needham Building in Cambridge, U.K. This invitation-only gathering drew scores of members of government, academia, and industry across Europe, in addition to Microsoft customers and partners, and members of the media. The next day, Herbert—son of parents who served in the Royal Air Force, married to Jane, and father of grown children Jonathan, Christopher, and Jessica— took a few minutes to discuss the event, his career, and the possibility of meeting Queen Elizabeth II during the honours investiture ceremony next month.

Q: How did your Enabling Innovation Through Research event go? Did you accomplish what you intended?

Andrew Herbert
Andrew Herbert

Herbert: I thought it went very well. We invite Microsoft subsidiaries around Europe to bring press, academic partners, and people from government so we can show them what Microsoft Research does and talk about the impact we have on Microsoft products. It’s helpful for the subsidiaries to know what some of the longer-range developments are, to show people that Microsoft is achieving innovative things, and to emphasize that Microsoft is doing R&D in Europe.

The last point is important: Europeans expect multinational companies to contribute to Europe’s economy directly and not just limit themselves to sales-and-marketing operations in the region. If a company actually engages in R&D in the Europe, it’s seen as a much stronger commitment than that by a company that just has people who sell things.

So the event is a blend of high-end technical marketing, communicating how innovative Microsoft is, showing what Microsoft Research is doing in Europe, providing a chance to meet some of our European talent, and talking about how we partner with universities in the region. It’s a broad spectrum of messages, but the key is innovation. If a company, as Microsoft does, has a world-class basic research lab, that generates fantastic ideas to feed into the innovation engine that’s the rest of the company.

Q: Did you have specific technological messages to deliver?

Herbert: I wanted to discuss why cloud computing is important and what some of the technical challenges are. I wanted to talk about natural user interfaces and the work we’re doing that supports that. The other topic was the growing use of inference, which is popping up everywhere: in Web search, computer games, natural language translation, and scientific analysis of large data collections.

If I had to say there was one underlying theme, it was data. I talked about data through the three lenses of cloud, natural user interface, and inference. Cloud computing is how you can process lots of data. If you don’t have a complete model of what you’re trying to do, inference is a great way of working statistically with data. And, obviously, if you’ve got lots of data, how do you access and interact with it easily? That’s where the natural user interface comes in.

Natural user interfaces are about how computers communicate with people. The computer, rather than being something we sit in front of and use as a tool to do something with, now stands next to us and is increasingly becoming more of an assistant and something that’s there to help us.

These topics are ones we talk about a lot in Microsoft Research, and we kind of take them for granted. When you talk to the wider audience about them, you realize these are fresh ideas, and people have lots of questions they want to have answered, which is where events like today’s feature.

Q: I understand that you are in your 35th year as a computer scientist.

Herbert: It’s scary, yes.

Q: When you reflect upon your career, what are the highlights?

Herbert: I was really fortunate. For people who started their research careers at the time I did, it was just as the personal computer was being invented. People like Butler Lampson, who is with Microsoft Research now, was at Xerox’s Palo Alto Research Center doing the early work that created the Alto and the Ethernet. I was at Cambridge University. We did a lot of work with personal computers, inventing ring-based local-area networks and processor banks, an early example of scalable computing. It was really exciting to be on the ground floor of so many new ways of thinking about computing.

A lot of things were just starting. People were realizing that writing in assembly code wasn’t the way of the future. It just wasn’t productive enough. High-level languages were just getting traction. High-level languages make you more productive and help you write code that is more likely to be correct. Formal specification was just being invented. The relational database was just about being invented. All these things were going in the late ’70s and early ’80s. It was a fantastic time to start doing research. You saw the foundations of modern computing being laid and got a chance to be part of it.

The other thing I think is interesting, reflecting back over 35 years of research in computer systems, is how in this field we go around and around in an ever-increasing circle. Fundamentally, the work is about exploring engineering tradeoffs. As new generations of computers and networks come along, we can construct different kinds of computer systems. Different aspects become relatively faster or slower than before, creating new challenges and opportunities. So we find ourselves grappling with familiar problems, but arriving at quite different solutions each time around. That’s really quite fascinating.

The irony is that, in some sense, I’m having the same debates with my researchers who work in systems today as I had with my first research students in the 1980s. Technology has moved on. The problems have gotten to be on a bigger scale, so they’re still fresh. But fundamentally, they are the same questions.

Q: How did you end up at Microsoft?

Herbert: I joined the company in 2001. It’s an amusing story. The Cambridge lab was founded in 1997 by Roger Needham. Roger had been my Ph.D. supervisor. I’d done my Ph.D., then stayed as a lecturer—that’s a “professor” in American terms—at Cambridge for about 10 years. Then I started my own company, which did contract applied research.

When Roger was creating the Microsoft lab, he called me up and said: “This is all terribly secret, but we’re planning a lab in Cambridge for Microsoft. Would you be interested in joining me?” It was frustrating, because he called me literally a couple of days after, with a bunch of colleagues, I had just raised funding for a startup. When you’ve done that and just built a management team, you can’t walk away from it. I had to say to Roger, “If you’d called me a week earlier, I could have said, ‘Yes.’” It sounded like the dream job, but, the startup was taking off, and it was the right thing to stay with.

The startup was successful. It was acquired by Citrix, and I spent a little bit of time with them as director of advanced development. Then, late in 2000, Roger called me and said: “Well, I’m still interested in you as an assistant director. Are you available now?” Yes, I was available and came into to Microsoft immediately. It’s not often you get offered that kind of job twice. I’m very pleased that Roger did come back a second time around.

In 2003, unfortunately, we lost Roger to cancer. That came as a great shock to all of us in Cambridge. When he hired me, he was looking to a future in which he would probably hand over the day-to-day management of the lab, take more of a chairman role, and still be engaged. Sadly, that never got to happen. If he could see where the lab has got to since those early days, I think he’d be really proud of what we’ve accomplished. It’s a shame he didn’t get to see our success.

Q: Has the lab evolved the way you expected when you became the managing director?

Herbert: In many ways the shape of the lab was already determined by the time I joined. There are certainly a number of things that I’ve done that have added to it. When I first came to the lab, we had three main areas: programming languages, machine learning and perception, and systems. One of the things Roger and I talked about was that we were under-represented in applications, user interfaces—that is, how people actually use technology.

That began to change when Roger and I hired Ken Wood. Ken built the Computer-Mediated Living group, and that brought people like ethnographers and hardware designers into the lab. These were really complementary skills to hard-core computer science and engineering and significantly broadened the scope of our work.

Another thing which certainly wasn’t on our radar when I joined the lab, but has turned out to be a great success, is having Stephen Emmott join us and start our work on computational science, using computer-science ideas to help advance the pace of scientific discovery. When I hired Stephen, he told me: “It’s not just computer scientists you should be talking to. You should be talking to other scientists, because they are building ever more complex models of physical and biological systems, and there might be exciting research to do that could help them and advance our own ideas.”

Another aspect we have strengthened from the early days of the lab is technology transfer, making a strategic contribution to the rest of the company. That’s a challenge when you’re a remote lab. I feel really pleased about our successes like F# and AdPredictor, to name just two.

There’s always a risk of an overseas lab being seen as a kind of ivory tower in a distant place. I wanted us to be an integral part of Microsoft, and I think we’ve achieved it.

Q: You recently added responsibilities for policy and engagement in Europe. How has that been going, and how have you been able to incorporate those duties into what no doubt is an already-busy schedule?

Herbert: [Smiles] I just redefined the day to be 36 hours long.

I was doing a fair amount of that work as part of my job anyway. I have asked Andrew Blake, Ken Wood, and Chris Bishop to take on more of the day-to-day management operation of the Cambridge lab to help me find more time, and they’re doing a great job. I certainly do a lot more traveling than before.

There always has been a pretty strong demand on the head of the Microsoft lab to join advisory boards and government consultations, because people are interested in Microsoft’s view of where technology is going. Governments want to understand innovation and innovation policy. What is the potential impact of new technology for industry and the economy? They have questions about what should be taught about computing in schools. Why is it so few kids want to study computing?

There are big questions on the policy agenda, as well: What role can information technology play in helping improve the environment, mitigating climate change, addressing energy issues? What about net neutrality? Can we make better and greater use of unlicensed radio spectra? Questions about Internet privacy and, as we get into cloud computing, questions about data sovereignty.

Those are big issues on which Microsoft has to speak.

In some sense, it’s been stepping up to a role that was already there, and it’s something I enjoy doing. For the last 10 years of my career, I have been involved with professional bodies and national academies, which is where many of these things get discussed, and I find that stimulating. It’s interesting to hear other people’s views. It’s instructive to debate the points and understand what the real issues are.

Q: Over the years, you’ve received numerous awards, including, recently, the OBE. How do you react when such awards are announced?

Herbert: I’m certainly pleased to receive them. It tells you that the community you’re a part of respects and values what you do. That is very encouraging. They invariably come as a great surprise, since the nominations are made by others, and it’s nice to be rewarded for things that I enjoy doing and feel passionate about.

The OBE was certainly a surprise. You get a wonderful letter—a very British letter—that sort of says: “If you were to let us know that you are minded to accept it, then we would probably be inclined to offer to you an OBE. Are you interested?” No one wants the embarrassment of either side saying no. So you reply back and say, “Well, yes, if you would offer it, I would probably accept it.” It’s very British, talking in that indirect way.

On June 9, I go to Buckingham Palace and receive my OBE. It’ll be a nice day out. There’s quite a formal ceremony. It takes place in the morning. There will be many hundreds of others collecting their awards. The ceremony will be led by one member of the royal family—you never quite know which, it could be the Queen, it could be the Duke of Edinburgh, it may be Prince Charles or Prince Andrew—who will say some words and hand me my award.

After the ceremony is a reception in the gardens of Buckingham Palace, which are rather nice. I’ve been there once before. So I might have a cup of tea with the Queen, if I’m lucky!

Q: What have you yet to accomplish?

Herbert: Engineering and technology are my passions, and there’s a management version of that—mentoring people, helping an organization excel. I really enjoy that, and the nice thing is that, as the lab has grown, as we’ve been asked to take on new responsibilities, it’s brought along new challenges for me to exercise those skills.

The technology-strategy and -policy work, that’s a new thing that’s come along, and that’s good fun. I’ve also been asked to look after the two innovation centers. We have one in Germany, and one in Egypt. They do applied research and advanced development. It’s great fun working with them.

I have a super time with the guys in EMIC [the European Microsoft Innovation Center, based in Aachen]. They are interested in embedded systems. We’ve strengthened the link with Windows Embedded and aligned what EMIC is doing with the new strategy that Windows Embedded is developing, and this is turning into a fantastic partnership.

The guys in Cairo [at the Cairo Microsoft Innovation Center] are doing super stuff with Arabic search features to make it easier to use. They’re also working on Arabic natural language on mobile phones for Egypt and the Middle East region. It’s great to visit them and learn about another culture.

New things keep coming along, and I enjoy a challenge. They’re invariably about something rooted in technology, working with a group of smart, motivated people and helping them do even better than they were doing by themselves. That’s what I enjoy.

Q: Finally, if you ever find yourself with free time, what do you like to do with it?

Herbert: Actually, I’m quite ruthless about making sure there’s some free time in my calendar, and a lot of that time gets spent flying historical aeroplanes. I have a machine of my own; it’s called a Chipmunk. The design comes from de Havilland in Canada; mine was built in the U.K. It was the training aircraft for the British Royal Air Force in the 1950s. It’s called the “poor man’s Spitfire,” because it’s cheaper to run than a Spitfire, but it’s an equally beautiful airplane to fly, and it does aerobatics very gracefully. “WZ879” is my pride and joy.

I also belong to a group that maintains two old biplanes that date from the 1930s and ’40s. The design actually goes back to the ’20s, so really they’re a First World War design. They’re great fun to fly on a day when it’s blue skies and sunny. I fly them in the winter, too, and that’s not necessarily such great fun—it can be quite cold at 5,000 feet in an open cockpit on a winter’s day!

I have a strong interest in steam railways—what you’d call railroads. I’ve got a large machine shop in the back garden of my house, and there’s a project there that is waiting for me: to retire one day and resume building a working, one-eighth-size model of a famous steam locomotive, a project I started many years ago.

My wife and I enjoy traveling and doing cultural things. My kids are grown-up and independent. I guess, one of these days, they’re going to produce grandkids, and that will keep me occupied, but, hopefully, not for a few more years yet.