Q&A: Microsoft chairman Bill Gates

Last week Microsoft announced the first details of Longhorn, the next version of the Windows operating system. Computing spoke to Microsoft chairman Bill Gates at the company's Professional Developer Conference in Los Angeles to discuss Longhorn and his vision for the future of the computing industry.

How do you see the next generation of hardware and software, both in the domestic and corporate environment? Do you think that Internet will be a transparent and fully integrated network with an operating system for the user?

The computing expanse will be integrated in more and more capabilities. That was our philosophy with the browser from the very beginning. We want to take that to a whole new level in terms of going out to get information, yet being able to do it in such a way that you know whether you're getting secure information, even as you go out into the Internet.

Until recently, the programming model didn't unify things; until we had the concept of web services, software on the Internet could not talk to other software on the Internet. The only thing that worked was to move TCP/IP or to put up screens in HTML, but software couldn't talk to software.

It's pretty fundamental to think about web services and how they are built in, and that's what really takes the Internet to the next level, where you can go out and get price quotes or the latest results or customer satisfaction.

What do you think developers can do to really harden their stance on security and what is Microsoft doing to help?

Well, you don't need a perfect code to avoid security problems. There are things we are doing to make code closer, in terms of appropriate tools, 'securiotics' and things like that.

However, for the user there have always been two key techniques: one is firewalling and the other is keeping software up to date. No problems will have happened to you if you have done either one of these things.

We can do very dramatic things in the short term, to do with firewalls and updating infrastructure. We've seen, in terms of magnitude, less vulnerability in the code that's been through the new tools. We've had about twelve issues in the eight-month period for Windows Server 2003, whereas with the equivalent level of tack in the previous generation we would have had over 100. We are talking about a factor 10 difference. We could get another factor 10 drop, which would get us down to 1.2, with improvements in other things. That's what people want but it doesn't happen overnight.

Longhorn includes many technical things that appeal to developers. Are you expecting software developers to take your lead and follow what Microsoft is demonstrating with Longhorn, or do you expect some real innovation to take place to take advantage of the platform?

The most popular application on Windows by a substantial amount is Microsoft Office and we are already hard at work on a version of Office that requires Longhorn.

That is what most of the Office team is doing now. Every major version of Windows, where we have really had the Office team engaged, requires deep involvement.

We always count on the Office team's involvement, but their really deep involvement comes when they are giving the full 100 per cent. This was the original set-up.

Excel was the first really deep application that ran on Windows - there were some before but they were not really that full-blown. Even things like printing and fonts, that we don't really think about today, got refined on that version.

So I certainly expect developers, whose names I don't know, to come along with new things to do with Longhorn. That's why with such a large event like this in the community, the global availability of all the developments and milestones along the way are so important to us.

How do you see the competition between Linux and Windows in the corporate environment?

Windows has the breadth of applications and the user understanding to make it THE operating system that people use on desktops.

As far as servers are concerned, there are a variety of roles that servers play. For the departmental server, Windows is overwhelmingly the strongest option. When you get to more central IT things like application servers, Unix has always been strong and we have seen a shift from expensive Unix boxes to Linux.

That's actually the biggest thing happening in Linux, the downfall of Solaris, HP/UX, AIX - those things are just going away as new design capabilities come up. People drop those high-end Unix.

We can come in with Windows and say "we can run those Unix apps, we can run those Unix utilities", but it's only natural for people to pick Linux as they make that migration. I would say that the majority of those migrations go to Linux. We get maybe 30-35% of them, by convincing them that, "hey, Windows is the better choice for that kind of migration". So, in those central IT scenarios they choose Linux because it is taking away from all the other pieces. There are only two operating systems that are really in growing in usage on the server, and they are Windows and Linux.

On the desktop there is only one thing that is significant and that is Windows.

There will be competition, but just because the kernel piece appears to be free, it doesn't mean that the cost of your overall project, or even just the software, is less.