Bill Gates: Unplugged

Microsoft's chairman compares Linux to Unix in the 1970s and urges people to "be careful" with utility computing.

Jai Singh Staff Writer, CNET News.com
Jai Singh is the founding editor and editor in chief of CNET News.com.
Jai Singh
7 min read
LAS VEGAS--For two decades, Bill Gates has used his Comdex keynote speech to mark out his vision for technologies from the Internet to XML. This year he used the bully pulpit to make it clear that the industry is at one of its perennial crossroads.

Once considered simple nuisances attending the digital lifestyle, cyberattacks and spam have morphed into disruptions costing millions of dollars in downtime and wasted manpower. Against a backdrop of mounting customer frustration with insecure digital infrastructures, Gates laid out his vision for a new era of technology that removes much of the hassle of being a computer user.

CNET News.com caught up with the Microsoft co-founder and chairman earlier this week to talk about the leadup to his "seamless computing" speech.

Q: You've been talking about seamless computing at this Comdex. Give us an overview of what's on your mind.
A: The key reason I picked the theme of seamless computing was to talk about the frontiers we still need to solve in the next few years. I see the things holding us back as being boundaries between different software systems...Why isn't e-commerce a reality? Why isn't managing your schedule digitally with friends and colleagues not a trivial thing to do? We can look and say that many of the problems are software challenges. Certainly, the solution to lowering operational costs on systems, the solutions to spam, the security challenges, the need to think of all these devices and how they work together--that's largely a software problem.

How can you hope to break down the seams if the vendors still don't really cooperate? Isn't that still a challenge?
Well, yeah. Take for example getting Microsoft Office and SAP to really work super well together. There's a Web services architecture

Get Up to Speed on...
Web services
Get the latest headlines and
company-specific news in our
expanded GUTS section.

that allows us to schematize these things. There's a pure architectural theme. In 2000, we committed ourselves to the .Net strategy. That assumed XML and Web services would become mainstream. Looking back, one of the things that was a clear success was the bet on XML (Extensible Markup Language) and Web services. People are just beginning to understand how profound they are as industry standards.

At the semantic level, we actually now have standards. That's been a holy grail for over 20 years. People spent a lot of time futzing around getting the bits to flow between machines and now that we have that, you think, "Well I can point a browser at any Web site. Why can't I do a query about all the sellers?" The reason you can't is because that's at a higher semantic level than just how to put the stuff on the screen. And it's far more complex. Only Web services give us a foundation for us to do that, so in a sense, a lot of the dreams of the '90s, like true e-commerce, had to wait for this industry standard infrastructure and the tools to be put in place.

What's your view of this idea of utility computing? And how does it speak to seamlessness if indeed this is a case of "here they go again," putting their twists and turns to what they want to propagate?
You have to be careful with utility computing. That was a rage during the 1990s, that everything would be hosted and moved outside the

Get Up to Speed on...
Utility computing
Get the latest headlines and
company-specific news in our
expanded GUTS section.

company. Where are those hosting companies now? Only a few things--like running Web sites--fit those models. The IT systems are your brain. If you take your brain and outsource it then any adaptability you want (becomes) a contract negotiation.

There's something common between the IBM message, the Sun message and the Microsoft message: Some of the things that you do with personnel to operate these systems today should be done automatically with software. We all agree it's a software breakthrough that will let people free up part of their IT budget that now goes toward operations and apply it toward new things. What's interesting is that everyone admits it's a software problem, not a hardware problem.

What's driving this? Is it marketing?
No. It's the development and operational and personnel costs that are really jumping up to be this huge percentage. You have to go after those to free up the most dollars for (IT) to innovate. We're not pro- or anti-outsourcing, but we think people have to be careful because there's certainly been more failures than successes.

Where does Linux fit into the picture?

Linux--which is only a kernel--is not where the interesting stuff is going on nowadays.
Linux is what Unix was in the 1970s: a perfectly reasonable operating system. There's a lot being done with it. Of the two operating systems that are gaining share, one is called Linux and one is called Windows. And Linux is wiping up Unix's share while our overall share has continued to grow somewhat. Clearly, at the server level there'll be intense competition where (companies) like IBM put WebSphere on top of Linux. That's very expensive and we just put the app server into Windows. And then people can do the value comparisons.

But with IBM pushing Linux, isn't there customer interest because they can offer a wider solution?
What do you mean by "wider?" Wider in terms of being more expensive? Wider in terms of the number of consultants required? What's wider mean in this case?

They're putting multibillions of dollars behind the Linux initiative.
But what does that mean? They don't develop (Linux). Now they charge people to run software on top of Linux and services.

Right. But now they're pushing Linux for the desktop.
People have had Linux on the desktop for a long time. It's not a substantial share of what's going on at the desktop.

So at the end of the day it's the Unix guys who are feeling the greater impact of Linux's adoption?
Remember, there are hundreds of incompatible versions of Linux. The fact that we call it Linux hides the fact that this driver works on

Get Up to Speed on...
Open source
Get the latest headlines and
company-specific news in our
expanded GUTS section.

this version, and this app works on that version. But there is a world where you don't test for binary compatibility because you don't have testers. It's just a different world than ours. That approach has certain benefits and advantages in terms of the way things get put together and ours has certain benefits and advantages. It's the primary operating system we'll be competing against.

Five years ago it would have been Windows versus OS/2. A few years before, it would have been Windows versus Macintosh. Before, maybe it would have been C/PM 86, and before that, maybe CP/M 80. There's always been some challenger to the operating system. Linux--which is only a kernel--is not where the interesting stuff is going on nowadays.

Talking about security, Microsoft's invested heavily in Trustworthy Computing the last couple years--including stopping all development to get the problems ironed out. How far along are you?
The initial wave of security problems were e-mail-related transmissions of malicious programs. Then a little over a year ago malicious programs spread in other ways as well. That was a big wake-up call because most of our customers didn't have problems with those things. It was the customers that had the up-to-date software and firewalls in place that didn't have the problems. For a high percentage of customers we hadn't made it easy enough for them to really audit that they had the right firewall capabilities in place and make it really easy for them to know which things they should pull through as fixes as opposed to feature improvements.

Our system is actually more robust because people are trying to do things to it.
There's been immense progress on this because it's been the top priority on what we're doing. Our system is actually more robust because people are trying to do things to it. In our case, we have lots of people seeking the glory of saying, "Hey, I found this flaw." They're not actually targeting a particular computer. So we're under very extreme scrutiny in terms of people trying to find these things. The big thing for customers is getting the updating and firewalls in place.

Why isn't the updating happening regularly? Does that aspect need to be automated?
Most of it comes back to us. Every patch has differences in terms of importance and the chance that it might disrupt things. If a patch is clearly labeled that it will cause no regression and you should put it in--you should just click and--boom! It goes in. But if something is a little bit of a new feature or a speedup that might disturb something, that has to be characterized in a very different way. And so you want clear labeling, clear regularity, and as few security critical things flowing through as possible. That's where we've made major advances.

During the negotiations to put the antitrust questions behind you, have Microsoft's product development plans been affected to the point that you feel you're fighting with one hand tied behind your back?
We've always tried to make sure that as we reach accommodations that our ability to innovate on behalf of consumers is not held back in some dramatic way. So far we feel good about our ability to innovate under the framework we're required to live in.