Remarks by Bill Gates, Chairman and Chief Software Architect, Microsoft Corporation
COMDEX Las Vegas 2003
Las Vegas, Nevada
November 16, 2003

BILL GATES: Well, good evening. It's great to be here at COMDEX and get a chance to look around and see what all the new things are. As was mentioned, I've been keynoting at this conference since 1983, and I went back and looked at my remarks from 1983 to see what was the continuity of all these different talks that I've given. And one of the slides that I actually remembered best was this one here, this is actually how this slide displayed. What happened was, it really literally was a slide. The word slide meant something back then, it meant something physical. And somehow my dad, I don't know what, it had gotten into the projector upside down. We also had a lot of vapor problems with the slides that year, a few of them were cracked. But even so, it was a message with the basic import consistent with what I've said every year I've been here since then.

This is one of the slides that I also used, talked about software, what is a great software package? In 1983, Multiplan was actually the best selling software package. 1-2-3 was just about to come out, for those who remember that. WordPerfect was just about to come out. And so a lot of amazing software that was built in the '80s, and that's the foundation we've been building on.

Some of the tough problems that were solved by those software packages are easy to take for granted today. Back then printer drivers were a bit problem. We didn't know how much help text to put in with the application because it would make it too many floppies for you to ship that program on. That was very tough. Of course, today, we can put almost infinite-sized tutorials with audio and video up on the Web, and have the product link into those.

But the basic principles here about having software that didn't surprise people, that you understood what it would do with you, that the commands would be understandable, that you could trust the program, those key attributes of software are still what we're focused on today. In fact, the idea of an undo, where if you do anything wrong you can go backwards, is pretty important, and not just for one application, for any set of systems working together, you'd like to be able to just set things back very easily, something that is being built into the software.

I talked in 1983 about software being the key factor that would bring the hardware to life, and realize a lot of dreams, including the dreams that Paul Allen and I had starting the company. Well, the successes of the '80s and '90s were pretty phenomenal, but those only brought us, I would say, half way to the dream of what software and hardware could do. Certainly, we achieved the title of the most empowering device that's ever been created, the best device for communications, the best device for letting you express your ideas, finding other people of similar interests. We are there on every one of those things. But in so many ways in terms of simplicity, and even in power, we're not there.

I think this decade is a very special decade in terms of what we'll be able to get done. I've referred to it many times as the Digital Decade, meaning that at the start of the decade the number of people who used computers for their regular activities both in their work and at home was quite modest. Creating documents, e-mail were really the only things that had begun to be used by hundreds of millions of people. By the end of the decade, we'll have over a half-billion people who have so many things that they do, whether it's scheduling, photos, music, organizing meetings, planning forecasts, all those things will be driven by software assists. And so there's more productivity to be gained in the advances that will come in the rest of this decade than the industry has delivered in our entire history up to this date. It's because we're building on the work that we've done before. And yet, the power is getting to such a high level that we can move even faster than ever before.

Now when we look at what's limiting us, my belief isthat the constraints that we have to get rid of now, the seams that hold us back, are of a different nature than it was in the previous two decades. The constraints now are much more about pure software challenges. In the 1980s, we were still held back by hardware. Moore's Law was necessary to get us a machine that could do graphical user interface. When we first put Windows up on a 286 machine, it only sort of was fast enough. It really took the 386, and some would say the 486, before graphics interface was clearly superior to what had come before that.

We needed hardware advances to let you run multiple applications and move data back and forth between them. So there was software invention, but we were constrained by hardware. In the '90s, we talked about the wonderful thing that would happen when these machines got connected together. And so at the start of the decade, the explosion of Ethernet started to fulfill that, the extra speed that the Ethernet took on, going from a megabit to 10 megabits to now often 100 megabits, and in some cases even a gigabit of performance, that wasn't there, although today we take it for granted.

And then, finally, the worldwide connectivity. And in the early '90s, if you'd said to somebody that you could sit at your computer and go get information from thousands and thousands of miles away and not even have it measured, so that you would pay some cost for that distance communication, that would have seemed very strange. The mindset was that communications was going to be very expense. And yet the model that succeeded was to not even measure the amount of network capacity that we're using, to just let you browse for a fixed amount. And that really revolutionized things. So, connecting all the hardware together, both locally and around the globe, that was the constraint we got rid of that let us do something very new and different.

Now, people thought that was it, that that alone was the last boundary, the last thing that had to be solved. In fact, that was wrong. When people talked about e-commerce, they didn't realize that representing the information of the buyers and sellers, and understanding the rich workflow that takes place in commerce, that a platform and a set of standards for that would have to be built, and that connectivity alone of the hardware was not enough. And so now, in this decade, we're engaged on delivering the final level of infrastructure, which is a software connecting infrastructure that connects all your different information together, that lets you work in a very natural way, connects you up to your speech, and your ink, and your photos, all those natural things. And so we talk about this as Seamless Computing -- the idea that we, through advanced software, will be able to eliminate those things.

So, what are these seams, what are these boundaries? Well, any of you who use many devices know that it's still quite an effort to get your schedule and files to move back and forth, and you run into mismatches where things like repeating appointments, or the spreadsheet file format isn't understood as you move back and forth. We'll have rights protection of this information. Again, that can be a real seam, that's a lot of work to move the information around. Many things, like using schedules, often isn't done simply because of that difficulty, and yet we are going to have many devices, from the wrist-sized device, to the pocket size, the Tablet, the desk, all of those are natural form factors, and in many cases you'll have multiple of them. Your desktop at work, your desktop at home, these should be connected. In fact, your e-mail, your personal e-mail and your work e-mail, you should be able to bring those together or separate those out -- that should be very, very simple.

If we looked at more business applications, we've got lots of them, boundaries between the productivity software and the structured ERP software, boundaries between corporations where the ease of moving information in a secure way is still way too difficult. And so by bridging these boundaries, by getting rid of these seams, the seams between the systems and the software, between the software and the software, between the software and the devices, we can deliver all of the scenarios that we've dreamed about since this industry got started, even to the incredible dreams of the 1990s that were not realized in the short time frame people expected they will be. The foundation for those things are what is being built, built in software, through various companies and industry standards that have just come together, particularly in this last year.

So, how do we make this progress? Well, it requires a lot of investment. It requires people building new software applications. It requires believing that these new end-to-end things can be achieved. It requires working with the universities that are often the first to see a pioneering opportunity, it requires the standards organizations to move forward into these new frontiers. Microsoft is very optimistic about this, and that's why we, for the first time ever, an IT company will spend $6.8 billion in R&D, that's double what we spent five years ago, and so it's been moving up at a pretty rapid level. That's what it takes to bring forth these advances we'll see. There were big bets behind this, the bet on XML. We focused on that even in 1997 as a way we thought these things should be built. And that's a standard that all systems are now implementing, and that data exchange is built around that.

Web Services, the next level above that, which is the protocols that let XML data move around, again, an industry-wide effort, very important. Managed code, a way of allowing you to write code that exposes you to less mistakes, and a very strong security architecture to have an explicit understanding of exactly what that code can and can't do and, therefore, have an expectation of how that information will be used.

So we really need these big breakthroughs. We need to push these things to this next frontier. And that's not going to be an overnight thing. Every layer has to build up. You've got to get the tools, you've got to get the pioneering people to build the applications. You have to have the end-to-end scenarios come together, different industries have to schematize their data in standard ways, but I see all of these things coming together.

The hardware is not holding us back, in fact, it's rushing ahead with some wonderful capabilities. The cost of the hardware to implement these systems will be very low. Even the natural form factors that we need for reading and tablets, and wrist-type devices, pocket devices, those are coming along at very full speed. The communications cost that we solved in the '90s continues to improve. So again there, if you look at an IT budget, that can be less and less of it. And we have to tackle the things that build in cost, developing applications, managing the networks, making sure that everything is running well. It's really that kind of advance that will make people confident to go and build these new applications. And that's, again, software that needs to do that.

Now, I get to spend a lot of time making sure that this drive towards these ambitious scenarios is really understood, and there's a strong commitment to it, all the way, throughout all the developers at Microsoft and in our partners, that build on top of the platform. I'm focused on making sure that strong commitment is really there, that people can see that vision. And I put together a little video that shows one way that working with one person at Microsoft I really did get the message across.

Let's take a look.

(Video segment.)

(Cheers, applause.)

BILL GATES: Well, we got Steve indoctrinated. (Laughter.)

I talked a year ago about some of these new devices and how they'd get connected together. One of the things that was featured was the idea of SPOT technology, low-cost chips that could receive information and change, things like a wristwatch or an alarm clock.

I've actually got my SPOT watch with me and it gets these FM signals that give me the sports scores, weather, stock, instant messages. This one is made by Fossil and these will be out in the marketplace early next year and so that's come to completion.

This is a device that's about four times as powerful as the IBM PC, sitting here on my wrist, and we download .NET byte code, managed code to this to enhance the applications whenever we want, things like showing a sports game in real time or whatever people come up with, it's a very flexible, programmable platform.

Also in this year we made a lot of progress in the pocket form factor. Of course, there's a trend in the phone devices to not just be voice but to also be very rich data devices, whether that means text messages, voice messages, e-mail, photos, ability to run software on these devices, and now we have great development tools, standard Visual Studio that can target these devices and do neat things.

This particular device, a super nice one made by Motorola; we've got Samsung making devices, many other people, and a lot of scenarios that are very important where you just automatically get your Outlook capabilities onto this device. Likewise, you get your Outlook calendar on this device, your lists, your contacts, those kinds of things.

And so we definitely have started down this path of Seamless Computing and devices working together.

Another good example is what we've done on Xbox where you can run a software title called Music Mixer that lets you get at all your music on the PC and lets you do very neat things using the Xbox there in the living room.

Another example is the Windows Media Center PC that brings in all of your audio and video capabilities and lets you have the TV guide and program recording customization, either through the normal PC interface or remote interface where you're just simply using a standard remote control to actually get at the power of that PC.

As we move forward on this there's been one clear message, and that is that we've really got to get the fundamentals right. Part of the fundamentals include what we talk about as Trustworthy Computing, all of the issues around security, reliability, manageability. And these are software problems. They're not easy software problems. To solve, for example, a security problem, some of the tough things that have sat there for over 20 years have to be advanced. The ability to examine a piece of code to see if it's correct, to take literally a hundred-thousand line device driver and prove that it doesn't do certain things that would damage the system; those are tools that we're not applying in our development process, the ability to scan and make sure that things like stack overrun, buffer overrun aren't going to be systematic problems that create software vulnerabilities.

The IT department needs a way that gives them the assuredness that their systems are going to be reliable, going to be secure and yet does that without them having to spend even the man hours on those systems they spend today.

The framework we're in is one where we've got to free up overhead in IT in order that those dollars can go against the new applications, go against Web service development, getting the wireless networks in, being ambitious about these new possibilities. And so it's quite a challenge to all of us in the industry to meet that expectation but that's something that's very doable.

The most acute thing here, of course, is security. It's been a key focus for us over the last couple of years, got kicked off just under two years ago, and certainly the largest thing that we're doing.

Now, to really provide security, the software has to be kept up to date and the software updates have to be clearly partitioned so that things that are just optional and new features are kept separate from the hopefully increasingly rare updates that relate to security issues that have really thoroughly been checked to make sure they won't cause any regression.

Having that capability and the ability to remove things, put things in, the ability to send them out in a very, very small form, the ability for consumers to have auto-updating capabilities on their system, those are the kinds of things we've been working towards.

We have two very key products that play a role here. One is our System Management Server, SMS 2003. This is a product that's about making sure you know exactly what software you have in your environment. This is a quote from one of our customers about how SMS has really transformed his ability to see what's going on in the network, to know what's going on with the operating system and all the different applications.

There have been significant changes in this, because this is a product that we've been iterating on as we've gotten more and more feedback.

Today over 90 percent of our enterprise customers have this licensed, and for managing their Windows environment, this is the solution. A few years ago there were many software products trying to work into this space, but this is the one that has overwhelmingly been chosen for this. And so it's got a huge responsibility to take on desktops, of which there are lots and lots, and provide patch compliance with very, very little time for people to spend on that.

Another key product that has to do with this idea of being able to create boundaries, create firewalls so that you never run into problems is called ISA Server 2004. This is a firewall type product that goes further than the classic firewall. In fact, we talk about it as an application layer of security. It can be used by itself where it will provide the traditional firewall capability in the application layer or a lot of people use this together with a traditional firewall and it simply takes and does the parsing of the software commands to understand exactly what's going on and what can be done.

Making it so people can audit do they have this thing set up the right way and making it very easy to do, that's a big challenge, but I think it's an absolutely necessary challenge to fulfill so that we can free up some of those dollars, get the confidence in these systems and have people charge ahead.

So I'd like to ask Zachary Gutt to come on out. He's going to give us a quick look at what we've done, what our customers asked us to do in these products. Welcome, Zachary.

ZACHARY GUTT: Thanks, Bill.

As Bill said, I'm here to show you two fundamental technologies from Microsoft that can help customers increase their security. First is SMS 2003 for patch management and second ISA Server 2004 to add defense and depth security at the network edge.

So let's start by talking about patch management. What we're looking at here is System Management Server 2003 and it's Microsoft's full-featured solution for inventory management, software distribution and patch management.

Now, to get some of that awareness I was talking about in my environment let's take a look at an SMS report. So today the patch in question is MS03-043 and we can see by this report that I have over a hundred thousand systems that are in my network that don't have this patch. Now, that puts me at about 1 percent compliance.

So now let me show you how we can use SMS to respond to this situation. You can see here that I'm already in the Distribute Software Update Wizard and this page is showing me all the patches that apply in my network. So I've already selected the one that I want to deploy and as I click Next, SMS is going to go out to Microsoft.com, download these patches, package them together and get them ready for distribution.

Now, I've always had the ability to target specific groups of machines to deploy patches to, but what's brand new in SMS is the ability to deploy patches during specific outage windows. Now, this is especially important for mission critical servers that are in data centers that can only be updated at scheduled down times.

And as important as data centers are, mobile clients are also a very key focus for SMS 2003.

So this is Bill's laptop here that we have on stage. Bill's been in Vegas for a couple of days but what he probably doesn't know is that the last time he connected to the corporate network the patches that he needed were automatically downloaded to his machine for him. So you can see here that I have a notification from SMS and it's telling me that I have about 25 minutes left before this patch is going to be installed automatically.

Now, Microsoft gives administrators the ability to have full control over the end-user experience and this allows them to give the end users flexibility to do things like postpone the patches if they're busy and need to install them later.

So the last big thing that our customers talked to us about was the ability to get up to the minute status on how a patch deployment is going, so we listened and this is also a brand new feature that's in SMS.

So looking here, we have a report that was generated about six hours into the patch deployment, and you can see here now that a number of machines are reporting install verified. What I want you to notice here is the richness of this data. You have machines in all kinds of states. Here we have some that are preliminary success, but I would like to focus your attention on this group here, no status. That's the group that this laptop here on stage falls into because it's offline. Now, the next time that it connects to the corporate network, when the patch has been installed, it will report install verified and the patch process will be complete.

So we can also slice this data a little bit differently. Here we're looking at a chart for my entire environment that shows how the patch deployment happened over time and you can see because my install was mandatory after four hours how the number of successful installations spiked right after that.

But, of course, we know that patching, while part of the answer, is not the full answer. So the other thing I'd like to show you today is ISA Server 2004 and how we can use that to add defense and depth to my network edge.

Now, this technology augments firewalls that are deployed today, traditional firewalls, by adding application layer security.

Now, I'm really excited to be showing this product in public for the very first time here at COMDEX.

So what we're looking at here is the first screen that an administrator sees after a successful installation, the Getting Started page, and it's going to lead me to define my network configuration as the first step.

So on the Network Configuration page I'm going to take advantage of a brand new feature called Network Templates that are going to help me drop this firewall into my environment very quickly and easily. I'm going to scroll down here and pick the back firewall template because I'm deploying this in addition to my traditional firewall.

It's going to start a wizard and on the first page it asks me to save my current configuration. Well, we don't have one right now so we'll move on.

Secondly here it's asking me to define my internal network IP addresses. You can see ISA Server 2004 has already done all the work and detected the settings automatically, so we'll move on.

The final page here is asking me to select a default firewall policy. Now, because I know that right off the bat, I probably have internal users that want to surf the Internet, I want to choose the restricted Web access policy here that's going to create these roles for me automatically. You can see down below that the rules that it's going to create are summarized here as well.

Now let's finish the wizard and take a look at what it generated. You can see here I have this rule for Web access that allows it from the internal network to the Internet for all users.

Now, a key thing is the ability to customize this rule for my specific environment. So let's say I wanted to add the ability to check your e-mail for my employees. Well, it's just as easy as flipping out this nifty context sensitive toolbox here, grabbing SMTP for e-mail, dragging it and dropping it on the rule, and it's just that easy for me to customize these rules to my liking.

So now let's say I have my network up and running, but I notice some of my employees are abusing this policy that I have for them. Let's say they're using peer-to-peer MP3 file sharing applications. Now, I don't want these on my network for several reasons: security, network bandwidth usage, and potential legal liability. Well, what you may not know about these applications is that to a traditional firewall they look the same as Web traffic.

So we're going to show you a brand new feature here in ISA Server 2004 called the HTTP Filter. This filter takes advantage of the power of application layer security to be able to detect this type of traffic and stop it while still allowing things like Web browsing to happen.

So I move over here to the Signatures tab and you can see I've created a signature for this specific peer-to-peer application. And this one I'm filtering on the user-agent header of an HTTP request for this specific signature. Now any traffic that matches this signature when this rule is enabled will be blocked.

You can see by looking at this list we can block many other different types of attacks with the same technology, Code Red and NIMDA here on the list, and with this add functionality it makes it very easy for me to add signatures to block attacks that might come in the future.

So in short I've shown you two big things here, first SMS 2003 and how it gives administrators the awareness of their environment and the ability to respond quickly and reliably when there's a problem; and second ISA Server 2004 about how it's easy to drop it in an existing network, easy to customize for my specific environment and how it has the application layer security capabilities that are necessary to detect the latest types of threats that I'm worried about.

Thank you. (Applause.)

Oh, by the way, I took care of patching your laptop for you, so it's all set when you're done with the keynote.

BILL GATES: All right, thanks.

There's a lot of activity in the security area. The industry over time will move for sensitive applications away from passwords to use smart cards. We'll move to improvement to the mail protocol that is enhanced SMTP where we'll be able to verify who mail comes from. Some of the weaknesses we have today go back to the very original standards of the Internet. Really knowing where patches are coming from and being able to identify that is again something that we're going to need to move up to the level of trustworthiness that we want for these applications.

Microsoft has a number of new things we're doing. We've got a free subset of SMS for people who want that for patching. We've got a way that we can actually turn the firewall on automatically for systems without causing compatibility problems, a lot of things that will move people in and make the default way systems are set up very secure and give IT a way of actually making sure that's what's going on.

Beyond security, another tough problem we've got today is spam. Here, the very benefit of e-mail that it's inexpensive to send messages is being exploited to try and find a very few set of people who actually want to respond to these messages. But because the cost is so low, even if one in 10,000 responds it's economic for them to send out that e-mail.

Now, there are several approaches being used to make sure spam doesn't become something that holds back people in using e-mail. One of those is the ability to detect spam mail versus legitimate mail, and we have an approach to that we call SmartScreen technology. It came out of our Research group when they noticed that the frequency of words and the types of links and things on the spam was generally quite different than normal mail.

But, of course, very interestingly the spam people, as soon as they saw that, they tried to evolve around it and make changes, but there, too, the unique nature of their mail was recognizable, putting in things like random words and hash codes and things that were an attempt to get around the thing that identified that that mail was absolutely the same.

So the SmartScreen is going to be in every mail thing we do. It's recently up in MSN and Hotmail. It's in Outlook. It's in a release of Exchange that we're making in the months ahead. So that's a very big step forward there.

We'll also let people have whitelist capability and then use proof that if somebody who's not on your whitelist, if they have certain proofs like that they've done a computation, or that they know somebody who trusts who they are, or that they're willing to put a monetary amount on the line, various approaches that will guarantee that even mail that comes in from outside from a stranger can get passed through appropriately.

David Coursey is one of the people who's been using the spam filter in Outlook and he's one of the many people who have commented that it's working very well, it really is making a dramatic difference there, which is, of course, super important.

We believe these new approaches will shift the tide, that between what we're doing with technology and what's being done on the legal front, it makes the business proposition for spammers no longer attractive. And we've got to keep working until we achieve that. And I believe very strongly, that's an achievable goal.

Let's move on and talk more about Seamless Computing, and where that comes into play. There's a lot of key boundaries that I talked about. There's the information you have in your personal setting, versus what you have at work. There's when you do an application, the design phase is very separate from the development phase, which is very separate from the deployment phase. If an error state needs to be programmed during that development phase, connecting up that error condition to the management system, so that right thing is triggered, that's very complex today, or taking information about what goes on when that application is running, and seeing it back through the view of these high level design tools that we're initially used, that's very difficult. So those are seams that we can get rid of to make application development a lot easier than it is today.

So as we sit down with customers and think through these different scenarios, we see things that require software boundaries to be broken down.

I've shown that some of the major Microsoft products are the products that need to evolve in this way. We'll be putting new features in that allow these platforms to play a key role in Seamless Computing. A good example of this is what's being done at the server. At the server, the idea that you don't have to worry about scalability problems, that's very, very important. The fact that you can set resource limits for different applications is important there. Certainly, the kind of hardware advance, moving up to 64-bit with the Itanium, and 64-bit Opteron-type approaches, those are very exciting things, because it just gets rid of limitations that have been there. Performance, file-system performance in the new server is about twice as good, because of optimization paths that we looked at.

In terms of developing applications, it needs to be as easy to build an application on a server as it is down on a desktop. There's advantages to both, and you ought to be able to use any code you write in both locations. And that was part of the vision of the .NET approach, to put the .NET runtime on the server, but then have that same approach on the client. So you could use applications offline, but also have them up on the server, where they can be administered, and they can have access to shared information.

The Web services approach needs to be built into these systems. It's not Microsoft's philosophy that you should have to buy an expensive application server to play in the world of Web services. That is going to be standard in the operating system, because the operating system itself, to manage its different components, to expose its state, it is going to use Web services. Part of my job as chief software architect is to make sure all the things we do are built, the components connect together, using these Web-service type approaches. And so that means that it's a fundamental element from the operating system on up.

For information workers, building in the idea that they can create Web sites in a very ad-hoc way, not have to go to IT, that they can set up the permissions, so if they want to organize people around a meeting they don't have to just use e-mail with attachments, or just use file server, they can set up a Web site, choose a template, they're off and running. And zero overhead for IT, because they can put up to 50,000 of those on a single server. The idea of rights management, that you'd be willing to put information in e-mail, or on a SharePoint Web site, and know that it can't be spread beyond the set of people that that information should get to, and that's very important, whether it's medical information, financial information, a lot of sensitivity to make sure that people don't inadvertently forward information in a way that isn't appropriate.

We recently, just in the last month, launched a new version of Office. And there are a few ways that this breaks down, some key seams that I think will be very important. Programming against Office, historically, was something that was very popular. A lot of people wrote Office macros. But, those Office macros forced you to think about the application and work through the verbs of the application. You didn't have a direct path into the data. And so, say you just wanted to update a table inside a word processing, or spreadsheet document, you had to understand how to navigate according to that application, so in Excel, selects, rows, columns, all those things. We've taken XML and said, look, let's provide a direct XML path in. So as long as there's any type of tags there on that document, you can read and write the information very directly. You don't need to know anything about the application. In fact, the way you do it is independent of which application is involved, whether it's Excel, or Word, or any of those different things.

So Office XP was a release that had features, like Outlook advances for the end user, it had One Note, for note taking in new ways, and it had lots for solutions, InfoPath with XML, the SharePoint and this new way of doing programs. We've had twice the volume of licensing, both in retail, and through our licensing agreements than we had for the previous version, which was Office XP. So actually adding all the online value added, the new modules, and these approaches has gone super well.

One way to get a sense of what customers are doing with this is to see an example of that. So I'd like to ask I think let's take a look at one of those. I'd like to ask Bobby Moore to come out and show us the new Office at work, using the magic of XML.

Welcome, Bobby.

BOBBY MOORE: Thank you, Bill.

Today I'm going to show you an example of how you can leverage Microsoft Visual Studio tools for the Microsoft Office System, and the new Tablet PC OS to build robust and powerful information worker applications over the Microsoft Office System. Now, specifically I'm going to show you an example of how Nationwide, which is a global insurance and financial service organization, leverages the Microsoft Office System, as well as these other tools to actually automate the internal process of building excuse me, of reviewing IT projects.

Now, Nationwide has recently decided to make the Tablet PC its standard computing mobile platform, for all information workers that actually need mobile PCs. And they're also piloting the latest version of Office inside their IT organization. So let's assume, for example, that I'm an information worker, and I'm actually a critical stakeholder in reviewing this process. You'll notice that I can actually see this project, in Word's reading mode, as a Word document. Again, this is an extremely intuitive way of actually reviewing this document, because Word is completely Tablet enabled, you'll notice it's actually very easy for me to actually annotate this document, right there within the document itself.

Now, when I'm done reviewing this document, I can actually bring up a survey that allows me to actually assess this project. Now, as I call this survey, what you're not going to notice, and I'll actually bring up the codes to show this, is that this survey is actually created in managed code, leveraging Microsoft Visual Studio tools for the Microsoft Office System. And this survey is actually created using C#, and this is a great example of leveraging managed code to actually drive the Office application.

Previously, you'd have to use Visual Basic for Applications, VBA, to actually drive the Office application, but now you can actually do this with the robustness of the Visual Studio.NET platform to actually drive your Office applications.

Now, let's actually return to the survey, and one of the things you're going to notice about this survey is it's completely Tablet enabled. You'll notice as I mark off the survey, this is actually going to drive the business logic behind how the overall quality assessment is created. And I'm actually going to suggest that I have a problem with sponsor agreement, and that's going to drive the overall quality to be poor on this survey. Now what I need to do is actually add notes, add a phone number, and my signature, and submit the survey. I'm going to do this using the Tablet PC.

The Tablet PC has been in the market for about a year now, and we've received great customer feedback. Some of the feedback suggests that our customers actually want an improved ink to text experience. So now let me show you what the Tablet PC has to offer. As I click inside the note tab, you'll notice that it's very easy for me to bring up the Tablet Info Panel. And in the past I would actually have to go down to the task bar to bring that up. So I'm just going to start inking a very quick note to myself, in actually very awful handwriting.

And one of the things you're going to notice is that the tip is actually growing as I need it to grow. And you'll also notice that I actually get real-time recognition, and not only do I get real-time recognition, but I can actually click inside of the tip, so if a mistake was made, it's actually really easy for me to actually alter that mistake inside of the Tablet input panel.

So, I'm going to go ahead and insert these notes, and one of the problems that I've always had with Internet phone numbers is, often the Tablet PC recognizes the numbers that I write to be letters. Well, now with context awareness, the Tablet is a little bit more intelligent about how it handles it, and it actually allows me to constrain the value that the Tablet looks for. So even with my awful handwriting, again, it's actually very easy for the Tablet to recognize that what I actually want to write is a phone number. And it does a very good job of recognizing that phone number. And if you'll notice, it's very easy for me to submit that phone number right inside of this survey.

I'm going to end by signing my name, and I'm actually going to add a little symbol that suggests that I'm a little upset with the status of the survey, hit the submit button, and Visual Studio Tools from Microsoft Office System is actually going to make a secure call to a Web service, and it's actually going to submit that information via that Web service.

So, now let's go into Excel and let's switch hats for a second, and let's suggest I'm the CIO of this organization. I want to see the complete view of all of my projects across the entire organization. Now, I'm going to draw your attention to Project Colombo, and you'll notice that the risk assessment here is that it's okay. Now, what Visual Studio Tools for the Microsoft Office System will allow me to do here is, I can actually hit the update button. It's going to make a secure call against that same Web site, it's going to bring down that information into this Excel spreadsheet, so as the CIO of this organization, I actually get the latest up-to-date information about how this project is progressing. You notice here I can just drill in, get more information about Project Colombo, I can see the security and IT value are OK, and I can see the little physical signatures of the individuals that have signed off on this project. I can actually even see the little symbol that I had just added previously to the form. And, again, as I want to get more information about why it's a high risk from a project management perspective, I can drill down even further and actually see the particular answers to that particular survey.

So, now, what have I shown you? Just a recap, this is an example of how the Microsoft Office System is now a robust programming platform, leveraging security, leveraging ease of deployment, leveraging all the robustness of Microsoft Visual Studio Tools for Microsoft Office Systems, as well as the Tablet PC platform that allows me to have extremely robust information worker applications.

Thank you.

(Applause.)

BILL GATES: Thanks, Bobby.

Well, in that demo, Bobby was using one of the second generation of Tablet hardware devices. Tablet has been out about a year. This is something that Microsoft is very excited to see the progress on. A lot of vertical applications, a lot of horizontal applications, and now close to a half a billion users of the tablet who are very, very committed users. We've had a lot of good feedback from those users, and are hard at work on a new version of Windows XP Tablet Edition. In fact, we'll have that by the middle of next year, and that's going to be a free product for everybody who has the Tablet today, it will install on the hardware they have, and they'll just get that. And it will include some pretty substantial advances. In fact, you saw the demo there, the high quality of the ink to text, and the whole correction interface which has moved ahead very, very dramatically.

In terms of the hardware, that's also advanced a great deal. We see on the screen five of the new models that are second generation models. A few of them are out now, the rest of them will be out in the months ahead. Toshiba, for example, has Portege M205 with a great docking solution there. We've also got there on the screen the new Acer machine, that also has a 14-inch screen. That was the first Tablet to ship with that 14-inch screen. Gateway has also a fantastic solution, again, a large 14-inch screen. And it's interesting, if you compare the notebook with the exact specs of this Tablet, the M205, the Tablet is just $100 more. A very powerful portable, with a DVD player and a large screen, and just a small premium to get all the additional capabilities. HP-Compaq has the Tablet PC, TC-1100. This has always been a form factor that has incredible versatility, the ability to work with the keyboard, or without the keyboard. And now they've added a lot of extra processing power. So the performance of the devices is dramatically better. Then we also have a very lightweight device, which is the Viewsonic V-1250, with a very first-class 12.1-inch LCD. So a lot of evolution taking place. We really thank the hardware partners that are driving forward on this, we thank the software people who have pulled these things together.

In the second year of Tablet we'll see substantial growth, as we move to new and different areas. A lot like graphics interface, where you get to an inflection point, and at some point knowledge workers say, hey, this is a tool for note taking, annotation and reading I'm going to have. And we're on the way down the learning curve that will get us in the years ahead to that point.

You've heard Microsoft talk about its commitment to Web services many, many times. It was central to our .NET strategy announcement in the year 2000, and this year was a great year for Web services. The standards being less at the basic level, and also the progress on the advanced Web services, which are things like security, reliable messaging, and transactions. We showed the .NET platform built into the operating system, nothing extra, interconnecting up with the IBM WebSphere software, and doing richer interoperability than we've ever had between a single vendor software. So that's phenomenal, to have schema driven information, being able to move back and forth, independent of the platform and the language of the applications on the two sides, in a very rich way.

Web services has taken off more inside companies than across company boundaries. That's about to change, but that was the easy bootstrap, where somebody had sources of data they wanted to connect within their own companies. Inside government it's been phenomenal. Only through Web services have we been able to take data out of existing applications, and pull them together, so a citizen who is interacting with a portal gets some coherency when they want to see the information, or change their address or something like that. So the number of examples in every industry, where the Web-services approach has led to new productivity, and really rich new applications is very, very large. What we've done now with Office and InfoPath just fit into this, and Visual Studio is the rich tool that we've got that lets people drive forward.

There was another big milestone for us this year and that had to do with starting the dialogue about what the next generation of Windows will look like. This, of course, is codenamed "Longhorn" and it was just this month that we had our Professional Developers Conference. To our surprise, this Professional Developers Conference had record attendance, more than the Win 32 Professional Developers Conference, more than the NT developer conference, more than the Internet-focused developer conference, every one of those major milestones for how developers think about tools and building inside the Microsoft environment. Conferences like that have had declining attendance, and yet here we were able to get more people than ever.

The reception was very strong, and even more importantly the feedback we've gotten is really helping us to refine what we're doing on "Longhorn."

"Longhorn" is a very ambitious piece of work. We're not even giving a timeframe. In fact, a lot of the people who got the prototype code that we handed out there said, "Hey, this looks pretty good," you know, "Jeez, you must be close to getting this done," and the fact is because of all the integration, the extreme high level of quality that we need to get, there's still quite a bit to be done and lots of room for people's input to really shape where we go, where we go with the user interface, where we go with the development tools.

There's a lot in "Longhorn" and the theme really fits into the Seamless Computing idea that I'm talking about tonight. There are seams as you sit down to your PC, the way you think of mail versus file, the way you think of browsing, the way you think of photos versus music, the way you think of annotating any one of those things or searching any one of those things; it's quite different. And yet if we had a powerful storage mechanism, which is what the Windows File System brings, we would change that.

Likewise, we want to bring visualization to a whole new level. That was one thing that people seeing the "Longhorn" prototype really were jazzed about, because it's so visual, it's so exciting to say, hey, this is a whole new level of taking the silicon magic and an approach to writing user interfaces that is way more efficient, separating code from on-screen behavior from actual on-screen appearance into ways that you can evolve your great designers, but developers can go and do their piece and you have complete flexibility between those things.

The actual runtime for Web services, very rich, enhanced Web-services capabilities, won't be an expensive piece of middleware. This is something that under the codename "Indigo" will be a fundamental part of the Windows operating system as we know it today.

So a lot to be done, a real journey that we're going down where we need to see neat applications. We had three or four third party applications that were early stage things shown there to give people a sense of what we've got there.

Now, the direction we're taking in all of our products is heavily influenced by the work that goes on in our research group. That is a really wonderful group with labs in Cambridge UK, at the Microsoft headquarters in the Seattle area and also in Beijing, China. And they get to play around with the very advanced ideas, make sure we know what's going on in the different universities and that we're really pushing the state-of-the-art.

One of the things that's influenced "Longhorn" is some of the idea of how we do information retrieval. So I'd like to ask one of the researchers, Susan Dumais, to come out and give us a look at some of the work she's done, because I think it really points in the direction we need to go. Welcome, Susan.

SUSAN DUMAIS: Thanks, Bill. Thanks for the opportunity to share with you some of the new technologies that we're exploring within Microsoft Research to facilitate people's ability to manage information.

Let me emphasize at the outset that the examples that I'll show you in a couple minutes are research prototypes. They're not shipping code. We work very closely with the folks in "Longhorn" to capitalize on some of the new technologies like WinFS to come to the shared vision of tremendously improving the ease with which people can manage tremendous amounts of information they're facing.

One of the key scenarios that we're targeting is information reuse. I'm sure everybody in the audience has been in this position. You're putting together a talk at the last minute, you're trying to find that critical quote, and you have no idea where you put it. It could be on your hard drive, along with 100-gig of your favorite material. It could be in the tens of thousands of e-mail messages that you have either on the local machine or on the server. Or it could be on a Web page you've just seen. You often have no idea where you put it. And in order to compensate for this people are spending a lot of time acting essentially as file clerks, trying to manage information re-access.

We think that the Stuff I've Seen prototyped is a very powerful alternative to that. What we've done in Stuff I've Seen is to provide unified access, cut across some of those seams that Bill was talking about, unified access to any information you see, whether you've encountered it as a Web page, an e-mail message or even a handwritten digital note.

In addition, when we have this infrastructure it allows us to then much more richly contextualize search; that is, make information available to users when they're in the midst of their job.

This next slide actually shows some of the kinds of information reuse patterns we've seen in some initial deployments of Stuff I've Seen. What you see here is that a lot of the information that people access is things that they've just dealt with in the last month or so, but at the same time people are accessing information in this case that's 16.5 years old. I don't think you used this to find your slide from the 1983 COMDEX demo, but people are using it over a tremendously long period of time.

So let me show you how Stuff I've Seen addresses information re-access at these different time scales.

What I'm going to do here is type in a query, "information overload," and we'll see what My Store has about that. You can see that was pretty zippy. This is running on my laptop behind stage. It's covering about 10 gig of data, maybe 100,000 items, and it returned 155 items that contained the words "information" and "overload." You can quickly sort by any column. We'll sort by date, path, whatever. You can quickly check any of these boxes to filter. Maybe I only want to look at information that I've seen in the last seven days. Maybe I want to omit e-mails. We think this kind of interface provides users with a really quick, flexible way to slice and dice the information that they're after.

The other thing to notice about this is that we're retrieving all sorts of information in the same place. We don't have to go to different retrieval applications to get in this case calendar appointments, Web pages, an e-mail message. You can see a .pdf file here. Here's a poster I was working on. This is one that's also very relevant to the demo that Bobby gave. This is a handwritten note about a talk that a colleague gave recently on knowledge-based approaches to information overload.

So we provided users with a unified access and a very fast and flexible interface to all sorts of information they've seen.

What I want to do is switch gears a little bit and talk about how a technology like this can help us better contextualize retrieval. Right now when you look for information you basically stop everything you're doing, pull up a separate application to look for information and then go back to the task at hand. What we'd like to do is much more seamlessly integrate information retrieval with other activities that you're doing. A very simple but effective way of doing that is this task bar that's constantly available. Regardless of what you're working on, information that you want is one click away.

Another way that we're looking at contextualizing retrieval is with something that we're calling Implicit Query or IQ for short. The idea here is we can look at any item -- in this case it's an e-mail. You can see this is an e-mail I received from a colleague earlier this week about a conference we're working on in Seattle in 2006. One of my jobs, as you see, is to be responsible for the keynote speaker and your name has come up, as you might imagine.

What the Implicit Query system is doing in the background is analyzing this text and proactively finding related information. Some of the information it brought up was the conference logo, which has a very nice Northwestern theme. So when I got this e-mail I knew that I had dealt with it, but I really didn't remember sort of what the outcome was. It found a reminder that I left for myself to in the beginning of 2005 send you e-mail reminding you that you said you would readdress this issue in 2005. You have to get on Bill's schedule pretty early for this.

So when I reply to this message -- OK, try again. (Laughter.) OK. (Applause.) Give me one more shot at this. OK. When I reply to this message what you can see is that the Implicit Query will automatically refresh -- okay, somebody is going to give me another shot at this. So let me respond to my colleague. "You'll never guess what; here I am on stage with BillG standing next to me." You can see that the information is changing as I'm doing this. "Let me ask him about this now." So if I can type correctly it will help.

So what you can see here is that in the context of my work the Implicit Query system is automatically retrieving information and returning what could be very relevant items without my ever having to explicitly issue a query.

So I hope that I've provided you a glimpse into the kinds of information retrieval technologies that we're exploring to provide both very fast and unified access, cutting across some of the seams you've talked about, as well as using background computations to help people find information that they might not think to explicitly ask about.

So thanks for your attention. (Applause.) Thanks, Bill.

BILL GATES: Thanks, Susan.

Well, that looks good. It doesn't look like she's going to forget about that conference that I need to speak at, but that should be great. (Laughter.)

Well, we've shown you some neat new things that really speak to what I think the frontiers of our industry are, standards and advances in software that can really give us all the scenarios that we've talked about for a long time, and being able to deliver those with tools not in the far distant future but simply in the years ahead to fulfill this vision of the digital decade.

There's a lot of work here for everyone. Customers have to weigh in on what the priorities are, making sure we strike the right balance on issues like the privacy and security issues. Competitors are going to come together to make sure that the standards are driving interoperability to this level, and I think we've done very well on that to date, but with still much more to be done.

But I hope the overall view here tonight gives you a sense that this industry really is just at the beginning, that the empowerment and advances are going to come faster than ever before.

Thank you. (Applause.)