MICROSOFT - A Cambridge too far?
When Microsoft set up its UK research lab in Cambridge, everyone thought the bogeyman was coming. A year on, things don't seem quite so bad.
This time last year, Margaret Beckett, president of the Board of thought the bogeyman was coming. A year on, things don't seem quite so bad. Trade, joined Microsoft chief technology officer Nathan Myhrvold and a gaggle of academics on stage at a central London hotel to officially announce one of the worst-kept secrets in the industry - Microsoft was going to Cambridge.
The prospect of a Microsoft research lab in Silicon Fen struck a combination of hope and fear into the area's computer entrepreneurs - hope that the vendor would lure new blood into Cambridge's closed technology community and turn Silicon Fen into the commercial success it had been threatening for years, but fear that the vendor would vacuum the 40 brightest talents from the surrounding companies, taking away irreplaceable software designers.
Satisfaction that Microsoft had given the city its blessing was tempered with scepticism. For a start, the vendor was following a trend, not leading it - Toshiba, Hitachi and the Olivetti and Oracle Research Lab were already there. Moreover, while Microsoft CEO Bill Gates forked out #20 million for a Winslow Homer painting, he dug only slightly deeper for European R&D - #50 million for the centre and #10 million for venture capitalist Hermann Hauser's Cambridge operation, Amadeus.
A year on, Microsoft Research is 14-strong. Hidden on the second floor of an anonymous building in downtown Cambridge, it's not immediately recognisable as belonging to the software giant - the owner is only apparent from the name on the door. There's plenty of room for expansion - the building will easily hold the 40 or 50 researchers that managing director Roger Needham wants by the end of next year. Apart from a few Microsoft boxes knocking about, the atmosphere is distinctly academic.
'We've got a flat structure here, no hierarchy,' explains Needham, pacing around his cavernous office. 'In a lab this size, it's how things ought to be done. We only hire researchers who are capable of driving themselves.'
Needham spends a lot of his time finding these researchers, trying to attract the big names in computer science to Cambridge to work on whatever they want to work on. 'There are no rules,' he says. 'We keep an ear to the ground for people who are discontented in their current jobs. Our challenge is simply to be of some benefit to the company. If Microsoft has someone who asks what to do all the time, it hired the wrong person.'
Microsoft Research UK is largely autonomous - Needham describes his status in the organisation as 'a bit like the governor of the Falkland Islands' - and there's no way the company will be hiring people to research a better taskbar.
Instead, research falls into less easily contained fields: security and insight into how to manage the complexity of areas such as electronic commerce. There's research into network performance, management of digital content and the future of computer books and programming languages. But we won't see the fruits of this work for years - if ever. Microsoft Research's assistant director Derek McAuley talks about a meeting with the NT development team, but the feature set has been frozen for months. Any of his insights can't be used until NT 6 at the earliest.
Microsoft Research's big brother in Redmond has had more effect on the way software works. Even the annoying little paper clip that pops up to offer help in Office 97 is based on Baysian Theory, a statistical method that lets Office 'guess' what you are trying to do. The Microsoft grammar checker is the first fruit of research into natural language by the Redmond lab. Eventually, Needham expects work done in the UK to filter into Microsoft products too.
Meanwhile, the rather neat office is about as far from the Xerox Parc hippy-filled brain tank as you can get. Needham is well aware of it, having spent some time at the Parc himself. 'This is the 1990s, not the 1970s,' says Needham. 'I thought beanbags were detestable and uncomfortable.'
He continues: 'People in computer research know where the action is and there's no doubt at the moment that Microsoft is where it is right now.'
Microsoft Research is very much the junior relation to the Redmond facility, where researchers can be counted in their hundreds. In its early days, users even mistook the UK lab for Microsoft's HQ and phoned it for support - some of Europe's finest thinkers were frequently being asked how to configure Windows. The facility has since gone ex-directory.
The UK lab has one main advantage - it works closely with Cambridge University's Computer Lab, which Needham used to run. Researchers give lectures to students and university staff are free to collaborate on what McAuley calls a 'shared work agenda'.
'Fundamentally, Microsoft is an intellectual property company,' McAuley says. 'But anything we do in this research work, we have the right to feed into our product groups.
'We're keen to make sure nobody gets exploited,' he adds quickly, admitting that during his former academic career, contact with commercial companies often left a nasty taste.
The third senior member of the Microsoft Research team is Chuck Thacker, who spent 13 years at Parc, where he was co-inventor of Ethernet and helped produce the first laser printer. Later, he led the development of the first Alpha architecture processor for Digital. He claims that despite a quarter of a century spent in research labs, development work is still exciting. And in the future, software will have to lead the way.
'The growth in hardware performance that we have seen will eventually run out,' he says. 'Light is too slow and atoms are too large. Therefore, we will have to get used to solving problems with a more limited hardware base.'
With that in mind, he is happy to commit to Microsoft, where budget is applied to solving theoretical problems rather than just fixing code. 'Microsoft is not interested in drawing back its research, making it advanced product development like a lot of other companies.
It's not doing that because senior management is technical and it understands.'
Microsoft is looking to the future and beyond. 'There are big goals here,' says McAuley, 'but there will be a lot of useful spin-offs along the way.'
This attitude contrasts with that seen by Needham and McAuley inside British universities. There, they say, home-grown talent is being stifled by a lack of funding and too much emphasis on short-term goals. Much of our native talent is being lost to the US.
'The government was forcing the message down our throats in higher education that we had to go and work with industry,' says McAuley. 'But when Microsoft offers partnership, some people say it's to rape and pillage instead. In truth, it is very rare to find an academic who can spare time for much hands-on research, so coming here makes you feel young again. It's ironic that British universities are being asked to do short-term research, while here we are being asked to look way into the future.'
Needham explains the Microsoft collaboration with Cambridge University is the first of many. There is already is some joint work on language with City University, a #100,000 investment in joint research with the pure mathematicians of the Dublin Institute for Advanced Studies and plans to collaborate with researchers in France and Switzerland. Overall, Microsoft Research is more keyed into the academic establishment than the Microsoft hierarchy.
Needham has a Microsoft Org-chart on his wall. It's like a family tree that traces the Microsoft hierarchy, but he admits he has only spoken to his manager once this year. And if people don't want to work for him because it means working for Microsoft, he's not about to try and do a hard sell on his employers. One of those reluctant researchers is his wife. 'She doesn't think much of Microsoft and she certainly wouldn't work for me,' he says.
Last year's announcement about the Cambridge lab started a rash of rumours that Microsoft was about to suck the talent from surrounding businesses, but there's little or no evidence that this will ever happen. Quite simply, the facility isn't commercial enough and it won't employ enough people to do that. What is probably much more significant to the city is the #10 million that Microsoft donated to the Amadeus fund. In the incestuous, closed world of Silicon Fen, #10 million doesn't go that far. But the effect, says Needham, has been to raise the fund's profile, which encourages other investors to back it too.
Eventually, Needham expects Microsoft Research staff to branch out and start up their own companies in the area. Then, he says, the advantage of a European research facility will become apparent.
Thacker, who was present during the growth of the real Silicon Valley, agrees. As venture capital is freed up, he says, British academics could become entrepreneurs just like their US cousins.
But in the next century, the race towards innovation will be far more competitive than it was in the 1970s and the talent will be even more thinly spread. McAuley, for one, doesn't think Microsoft Research is doomed to become a provider of grammar checkers and talking paper clips for Microsoft's overworked coders. And if the easy-going team at the labs is in agreement on one thing, it's that the work they do will yield some fundamental advances in the way we use computers.
In fact, McAuley says, it has an advantage compared with Parc. 'One of the biggest differences between Microsoft Research and Xerox Parc is that we have a product division that's interested in what we do,' he says.
It's ironic that the company that survived by buying technology rather than writing it's funding the generation-after-next of computer research.
One of the reasons has to be that for a company whose profits are measured in the billions, #50 million for Microsoft Research comes cheap. But in Cambridge, among the cash-strapped academics who don't have to apply for research grants any more, no one is complaining.
So what is the ultimate goal for Microsoft Research? 'The jackpot for us would be inventing a new business for Microsoft to be in,' says Needham.
'If you're selling sausage rolls to everyone in Cambridge, you need to think about baking something else.'
LAB QUOTES - WHAT THEY SAID A YEAR AGO
'We're here to conceive technologies that will lead to new ways of helping people. We don't know what they will be, but we have to remember that innovations we take for granted, such as the mouse, were the result of research undertaken 20 years before they came into widespread use.
Now is the time to be projecting well into the next century.'
Roger Needham, managing director Microsoft Research UK
'We're basically here because Microsoft faces a fundamental dilemma.
In order to make computers evolve, we need to make substantial investment in research and development. I think this is a good opportunity to reach outside the US to Europe.'
Nathan Myhrvold, Microsoft chief technology officer
'This really reminds me of the vitality of Silicon Valley in the early 1980s.'
Tim Negris, president and chief executive of Omnis Software
'In the 1980s, people got excited about Cambridge and the hype wasn't fulfilled. But in the 1990s, venture capitalists are getting returns of between 40 per cent and 70 per cent a year.'
Hermann Hauser, venture capitalist
'Most local people in Cambridge are graduates of the University. They are not taken in by Microsoft-rules-the-world arguments.'
Stephen Ellwood, hardware manager at ATML
'This is a real declaration of confidence in Britain.'
Margaret Beckett, president of the Board of Trade.