Microsoft finally gets its man

Among a crowd of eager photographers dressed as if for a combat zone, and well-coiffed public relations people in thrall to Microsoft, Roger Needham stood out rather clearly at a London hotel last week.

Then again, so would anybody wearing a yellow tie with a black shirt. But that wasn't all. Needham, 62, was noticeably calm and self-assured amidst the eager desire to please that was visible from Cambridge's vice- chancellor Alec Broers and the president of the Board of Trade, Margaret Beckett.

Microsoft, that supremely wealthy and powerful guest, had said it was setting up a research lab in Cambridge that would be "attached" (but not part of) the university there. It would invest pounds 50m over the next five years to recruit excellent computer researchers from around Europe, and students and staff at the university would benefit from the interaction. But, in fact, while Mrs Beckett and Mr Broers were proclaiming this industrial catch, Microsoft was delighted with having caught Professor Needham.

It's seven years since it first tried to lure him to head its research organisation at its base in Redmond, near Seattle. The fact that it has kept on trying to get him shows that as far as the biggest software company in the world is concerned, Professor Needham stands head and shoulders above the rest of the world's computer researchers. And if this computer Mohammed wouldn't come to the mountain, why, the Microsoft mountain would just have to come and set up shop where its Mohammed lives. Nathan Myhrvold, Microsoft's chief technology officer, paid lip-service to the idea that the corporation had looked at various other sites to locate in. But once they found that Needham had no other commercial ties, every other idea - including Boston, home of the world-famous MIT Media Lab and the digital guru Nicholas Negroponte, went by the board.

So what is it that Roger Needham has done, and is doing, that is so special? For a start, his history in the field goes back more than four decades. A potted history of his work?

"In 1949 we [at Cambridge] had the first modern digital computer," he recounts. "A couple of years later Maurice Wilkes developed microprogramming for microprocessors." Such "microcode" is the essential underpinning for all operating systems. And at that time there was no such thing as a microprocessor.

"In the mid-Sixties we made the first generally available timesharing operating system on this side of the Atlantic. In 1966 I developed a one- way system for encrypting a file: that's used all the time now by computers for password files.

"In the Seventies we were working on LANs [Large Area Networks], using the Cambridge ring." At least the name must have come easily, one suspects. "Of course, the Xerox Ethernet won out there. But we made a considerable contribution. And we also wrote a natural language and information retrieval grating formula [used for sifting data] with City University, which is now used by commercial companies for examining marketing data.

"Then, in the Eighties, we took part in an experiment linking LANs by satellite. And in the Nineties we have got what I think is the strongest group in security-related subjects in any UK university."

His own specialism is cryptography, which, with the advent of electronic commerce, is sure to be increasingly important (though Myhrvold denied that the UK was chosen as a base to avoid the US's export licensing laws on cryptographic software). And he retains the enthusiasm and vision for the potential of computers that mark out brilliant researchers from the merely excellent.

"One of the biggest challenges we will all face will be the management of digital content," he mused. "All content, whether films or books or whatever, is becoming digital, and can be rendered that way. Once it is, it will be amenable to management by computer. Except we don't know how to do that. We want to be able in the future to say, `Find me the film in the library which has that scene where the girl in red goes off on the right on a white horse.' Computers can't do that today, but they will in the future."

He finds such issues interesting, but Microsoft - and every other big software company - finds them desperately important. "Microsoft faces a fundamental dilemma," Myhrvold explained. "Our products must evolve. Every one of our products will be obsolete in a couple of years if we don't pump new technology into them." (Microsoft-haters will have their own version of his analysis of obsolescence.)

"Computers are too inflexible," Myhrvold continued. "They can't see, hear or understand us. To make them evolve as a tool we need to invent new technologies."

Which is just what they want Professor Needham to co-ordinate via Cambridge University's computing laboratory (with, of course, its own Web page at www.cl.cam.ac.uk).

The Cambridge lab has in fact already played a significant, yet unacknowledged part in Microsoft's evolution, getting it to expand the desktop of Windows 95 to include the Internet - rather than treating the global network as a threat.

Says who? Says the latest instalment of the Microsoft story, Overdrive: Bill Gates and the Race to Control Cyberspace, by James Wallace (a follow- up to his book Hard Drive). The date is December 1993, and Rob Glaser is in charge of investigating new technologies for Microsoft. A team meeting is called: he wants to show them a coffee-pot.

"The coffee-pot was just one of the many `cool' things on the World Wide Web that Glaser had wanted to show the team. In England, someone had rigged a camera to take pictures of a coffee-pot, and the live image was transmitted to a website." As it happens, that coffee-pot was in the computer lab at Cambridge. (You can find it there still, at www.cl.cam.ac.uk/coffee/coffee.html.)

The laboratory currently has research groups working in artificial intelligence, "automated reasoning" (which attempts to "prove" mathematical theorems by computer), natural languages, security, graphics, computer vision, and theory and semantics of programming languages - which could provide the keys of future mini-languages such as Java.

But one of the provisos about understanding Professor Needham's work is that what he is doing today will take a long, long time to arrive on your desk. Better to plan on its arriving on your children's desks.

It is always a startling experience to visit a computer research laboratory. As Arthur C Clarke has so famously said, any sufficiently advanced technology is indistinguishable from magic. And to go inside the MIT Media Lab or the Cambridge Computer Laboratory is to feel that you're watching machines work magic. You see maps with layers that you can navigate through. Systems that can solve theorems, or turn a sea of data into a recognisable set of shapes lying on a virtual plain, so that you can see their relationship to each other.

Those are on show in laboratories now, but don't expect to see them soon. "The pipeline is long," Professor Needham says. "Look at the mouse: that was invented 20 years ago, and it took that long to move from the laboratories to being ubiquitous.

"And an awful lot of research is viewed at the time as being futile. You know, with time-sharing we did a demonstration in 1968 with MIT using a transatlantic Telex line, showing how you could use the same system at such a distance. And the pundits said, `It won't work, it won't be economic, and people won't want to work that way anyhow.'"

And so now, of course, the pundits have all been proved wrong? "Er," says the professor, adjusting his glasses and for a moment looking like Eric Morecambe. "Actually, now it's obsolete. But it was the dominant form of computing for some time." Before it was wiped out by the PC - proof, if it were needed, that in the computer world, obsolescence is never farther away than the next round of technologyn