Remarks by Bill Gates
Most Valuable Professional Summit
Redmond, Washington
Feb. 11, 2003

BILL GATES: All right. (Applause.) I wanted to start by thanking you all for being here. It's a fantastic thing for the company, and all the product groups are very excited to get your feedback. We're trying to make the energy they've put into these things and the process, really getting that to reflect in new things we do in the product work even better and better. And the time you're taking by coming here - the passion you show about helping us move these things forward - is really quite fantastic and so I want to thank you for all the time you've put into it and I want you to understand how committed we are to taking the valuable learning and really taking full advantage of that. So thanks very much. (Applause.)

Next, I want to share some thoughts about where our product strategy is going and then I'll leave most of the time for answering whatever questions people have.

Microsoft in this period is the most optimistic company in the technology business. (Laughter.) That wasn't true three or four years ago. In fact, we found ourselves a little bit in a strange position where a lot of the hype was talking about things happening overnight and banking and shopping and all these things would be immediately different. And a lot of the focus was on the latest person to do a new Web site and the whole startup phenomenon around that and the valuations that went with it.

During that time a lot of the things that are hard and take time -- like handwriting recognition or model based tools or getting code quality up to a whole new level, unifying the storage -- you know, things that take six, seven year type technology commitments we were hard at work on and we knew that in the right time those things would come out into the marketplace, but we felt as a whole people were expecting more out of the changes that would take place, certainly in terms of the timeframes that were predicted.

At this point, most of those excesses are gone and you've have to say that to some degree there is an overreaction to that, many people saying, well, the dreams of e-commerce, of online communication, of paperless type transactions, that they're almost discredited in a way that is equally wrong, because every one of those things will happen. They'll take time in terms of the social acceptance, the price of the devices, the ease of use of the devices, the quality of the experience, the trust that people have in these infrastructures. These are not things that ever will happen overnight. And yet if you take a reasonable timeframe like the rest of this decade you will see dramatic differences.

At the start of the decade the only activities that were primarily done with computer help in a digital form were document creation and editing and electronic mail. By the end of the decade we'll add several dozen things to that list, things relating to note taking and reading, things related to bill paying, certainly everything to do with photos, everything to do with music, everything to do with organizing schedules. Those things will become mainstream in a way that it won't be novel at all that you'll expect those things to work in that fashion.

Some of the claims we make here are still considered extreme. Certainly the claim about reading, the claim about the device being smart and understanding what kind of alerts and interruptions are worthwhile to you -- whether it's a phone call or an instant message or an e-mail -- and working on your behalf to take the things that you never want to spend time on and not making those come to your attention at all and taking things you want to see in their right priority, bringing those to you on the right device at the right time, that's what we call the Information Age vision.

There's a lot of invention still to be done to make the Digital Decade happen. I'll talk in particular about all of the invention around Trustworthy Computing, bringing the reliability and privacy expectations of how all this works to a whole new level. Those are requirements for all these scenarios to take place.

But we believe in these scenarios. We know that people in the hardware business are continuing to make the advances that drive this forward. If chip improvements stopped happening, then the technology industry would be in a place where people should not be optimistic. Gordon Moore just yesterday at the semiconductor conference said that his so-called law he sees maintaining its pace for the next decade, and so that's a piece that fuels these advances.

The disk industry will continue to double storage size every two years. The optic fiber people, despite all the challenges they face, will continue to double the bandwidth that can go down a single fiber every couple of years. Flat screen technology, 802.11 and its variants providing us wireless connectivity; all of these things lay the foundation to allow far more scenarios.

And so we have to have the hardware magic, which fortunately Microsoft doesn't have to do that itself. We can count on our partners to do that, and then you need software magic, which is partly in the platform and partly in the third party applications that get built on that, and there we have a very fundamental role in terms of investing in the platform itself and getting the information out to third parties so they can build the applications on top of it.

We are putting our money where our mouth is on this one. We're increasing our R&D again in a fairly substantial way. And that's not because we see some economic turnaround in the next two or three years. We're not experts in that at all. We're doing this because we see the opportunity of better software technology, of a better platform enabling new things.

That's been a long journey to get to where we are today. It goes back to myself and Paul Allen dreaming about software, dreaming about a microprocessor that if you run Moore's Law in reverse for 28 years imagine how little power that 8008 had. It could almost do nothing. And yet we thought with exponential improvements that you might actually get somewhere.

So we've put together a little video that captures some of the twists and turns about what's gone on behind the technology, so let's go ahead and take a look at that.

(Video segment.)

(Laughter, applause.)

BILL GATES: I'd say this is the first audience that's understood all the subtle humor in that video. (Laughter.) Even Microsoft Bob people knew what we were talking about. That's pretty good.

This last year we've had a number of new things that had moved forward the foundation for how we see these new scenarios. Windows XP itself has been a very important thing for us, getting Windows to run a strong code base and building in for the first time the feedback mechanism, things like the crash analysis capability where we're getting literally on a daily basis a full log of all the reports people send to us to understand on our own code and in third party drivers or hardware where the various pieces of the PC ecosystem aren't coming together in a perfect way. That's really driving a lot of the breakthroughs we see in our tools, in our reviews and in how we work together with other participants in the ecosystem.

We got out some new versions of Windows -- the embedded Windows that's used in a lot of different devices -- the Media Center that brings that PC into the living room for video and audio experiences at a distance, not just the close-up screen but the remote controlled screen working off the same rich power, the same photos and music, and the Tablet PC.

When you think about Tablet, think about during the course of the day how can software help people as much as possible when they're in meetings, when they're looking at magazines, when they want to share their notes, and the Tablet PC is a real step up in how you can use digital approaches.

Here at Microsoft, of course, we've got the wireless networks installed everywhere and so in a typical meeting now we have a protocol: Is this a meeting where you can use your Tablet and just take notes or can you use your Tablet and also browse other things like e-mail that might be coming in. And for the main participants, it's often a focused meeting. But people, you know, the topic's not focused on them -- they can look at other things and still be getting a lot done while they're there when their particular topic comes up.

So tools like this when used in the right way really are significant advances in productivity. Anyone who's had a Tablet for more than a couple of weeks, you can tell them by asking them if they'd give it up, and it's kind of like graphics interface was, where they're completely wedded to that new approach that it lets them take.

Now in the home environment, you can say it's definitely a glass half full. Over 60 percent of U.S. households have a PC but the breadth of things people use that PC for is fairly limited today. It's a little bit of homework, some work that comes home, a few games but it's not a broad set of applications. It's not scenarios that involve the entire home.

And we need to make it easier to try out different applications to use for special things and to think of this as a platform where all sorts of entertainment -- recording TV shows and being able to play those back, replicating media out to handheld devices or, say, to storage in your car. Your car pulls in the garage, it's within the range of the 802.11 network so that your tunes and shows and things that you're interested in are automatically just sent out there. And so then it's connecting with all the different devices: the TVs, the PDAs, multiple PCs working together in a different way. There's a lot we can do on the home front. Media Center is just a start there in terms of the way we see things happening.

One of the key areas that we've been investing in very heavily is the idea of how you interact with the computer. The keyboard, the mouse, you know, those are the established means, but over time -- ink as a data type, whether recognized or unrecognized, voice as a data type, again whether simply recorded or recognized, gestures where you have video recognition that can understand who's in the room, what kinds of things are going on and whatever commands you give are able to be interpreted by the system -- these things will come into play because after all, with the miracle of hardware, the cost of having a camera there or having a digitizer just gets lower and lower, and if we can couple that to software magic, then we can deliver on scenarios that people really care about a great deal. So making this a better and better tool we are not even halfway towards that.

One of the biggest products we've ever had, of course, is Microsoft Office. We're on the verge of a major new version, which is the Office 11 release. That is both an advance in all the things that are internal, the modules we've had with things like XML or a whole new way that Outlook deals with connectivity, which is a huge advance. I encourage people who haven't had a chance to look at the beta on Outlook 11 to just even try that one piece and understand what a major change it is in terms of how you think about e-mail. Its ability to not distinguish between connected versus disconnected is a very big thing.

The two new modules pioneer new territory. OneNote goes beyond the journal application that's built in the Tablet, can let you do lots of outlining things. It also records the audio so that if you go to your notes you can listen to what was being said at the meeting when you took that note.

InfoPath is part of the XML revolution that is being reflected across all Microsoft products. But in order for the XML revolution to happen, one piece of it has to be rich viewing: the ability to create rich schemas and have right user interaction with XML documents; and that's InfoPath. Being part of the offering really provides that critical piece.

The bet we've made on XML and the Web services protocols that are used to exchange XML data is a bet the company strategy. We announced this almost three years ago with our .NET strategy. .NET is easy to think about. It's Microsoft's implementation of these XML Web services. And so there's many things we've had to do to get this going. One is to have a very advanced development tool that supports it. Visual Studio .NET was the kickoff there.

Another is to have standards, industry wide standards so that these protocols can be used not just from people who use Microsoft but from people who use IBM or other software as well, and that's why the Web Services Interoperability Group was created, sometimes called WS-I. And it's been great to see the participation there and the speed at which the specifications have come out and become very rich and even implementations of those are coming along in a very, very strong way.

So the bet we made several years ago is really paying off, and customers are seeing how this solves not just the business-to-business exchange of information, but even in a very profound way the exchange of information inside a company.

We're using this architecture ourselves to take on problems that have been very tough to solve in the past, things like managing desktop machines or managing many services. What is the infrastructure we use to do that? Well, it's XML Web services, the ability to discover things, to distribute events around things, to store things in a very rich format. This architecture is key to interoperability at many, many different levels, and that's why it's an architecture that I'm driving into all the different product group work.

The profound effect on Windows, on SQL Server and Office is hard to overstate. Office needs to understand XML in a native way. SQL needs to not -- like all databases -- convert things into tabular form and back to XML, but rather store the XML natively, and that's what the next major release of SQL is aimed at doing.

So, using this architecture, we don't just get bit connectivity, we don't just get TCPIP: it is everywhere, we get information that we can understand. So, this allows people who are working together a better exchange of data. It allows the impedance between the knowledge workers and the back-end systems to be far less between the various back-end systems, between the IT systems like the directories and the software updating things and the monitoring things they need to do and even between the devices, checking the state of the device, replicating information out to the device; all of those are now built on this Web services' approach.

So it's as much of a profound change as was the commitment we made to graphic user interface in the late '80s, or the commitment to Internet and all the browsing capabilities around that in the late '90s. And in this case, things are being driven forward, I would say, at an even faster pace. And so you'll see .NET not just in a few products but across all the things we're doing.

The customer reaction to this has been very good. As they see the concrete scenarios, they see the ability to not write so much glue code, and you had to have deeper insights into the information in their company; it's been very strong.

Now, we have to couple it with the fact that this infrastructure will reduce IT overhead as well. It will let the systems be more manageable at the same time they deliver more value. And in this era where people are looking not to expand their IT budgets but to get more out of them, that dual capability is a very important part of why we think even in this type of time period there will be a quick uptake of these new approaches.

A key element that I mentioned at the outset that has to do with the maturity of this infrastructure is that its fundamental reliability is at the level of other infrastructure systems -- say, the electrical systems, water systems -- and software-driven systems don't achieve that today.

There are some very key advances taking place. For example, in the area of redundancy, the basic approach of being able to have multiple servers handling a task, and that if any one of those servers fails, that the changes are transacted in such a way that other servers can take over without their being any loss of data. That approach, which was a special variant to the computer industry that Tandem & Stratus did many years ago, that's now a mainstream thing. And so, instead of being a variant with special development tools and approaches it's built in to every copy of the Windows Server.

The idea of how we know that people are authenticated, and not just use passwords -- the move towards smart cards and biometrics -- that's just beginning but there's been very good progress on that. Corporations are starting to understand how important that is.

Being able to have Web sites talk about what they do with information, talk about privacy, there's very good progress there.

Being able to have e-mail be authenticated so you know that it's legitimate and it hasn't been tampered with, the kind of domain authentication and user PK [Public Key] infrastructure, we need to make that possible. That's happening.

The ability to update systems very rapidly so that if any exposure is found in a piece of software you can immediately go out there and patch all those different servers, and yet that's done in such a way that only critical updates go across that. And so, people are willing to let their systems be very up to date -- in that sense, that is a key advance. It's something that we took on about nine months ago. Just recently, with the so-called Slammer attack, it was reinforced how important that is.

We've got to have the ability to get the latest improvements to have 95 percent connection, particularly in Internet facing systems, so that you don't have an environment where things can cascade.

There are a lot of industry joint efforts and standards necessary in this trustworthy area, a lot of work with government: things about spam laws and privacy laws and helping people to audit systems and having government systems be showcases of those things.

This next year is a pretty big one for us again. Last year I thought was pretty big with Tablet, Media Center, XP getting out in a very broad way. This year we have a server release, Windows 2003 Server. We've got the new Office version and we've got all sorts of evolution in the things around .NET and how developers are working around those things.

So with all that R&D investment that we're making you can see that actually something does result from all of that.

A new thing that we've just been talking about publicly that we were incubating for the last three years is called Smart Personal Object Technology and this is the idea that there are many natural form factors, and really five. There's the wall-sized device like this projector here or the big shared TV in the living room. There's the desktop screen, which will get bigger to give you more field of view as the price of screens comes down. There's the Tablet form factor that you can carry with you that literally is like the size of a tablet. There's a pocket-sized device that includes both phone and PDA, and those are kind of coming together. In fact, the handheld game, camera, phone -- all those things through the magic of hardware and software you will get in one very powerful device, and there are many people participating in that category.

But it's our view there's a form factor even one step smaller than that and that's glanceable information, where you look at an alarm clock or you look at a wristwatch or you look at a little magnet on a refrigerator, and you see information that you care about. And that's what we've created here with the SPOT type concept.

It involves some chip work that we're doing with National Semiconductor. The power of the chips that are going into this wristwatch are about four times the power of the original IBM PC. And it's kind of amazing to think that what's happening in this watch is we use FM sub carrier, and we download .NET byte code so you can program things up and it goes into that watch. That watch has 512k of ROM, 384k of RAM and runs at about 28 megahertz. So that's compared to probably the first PCs that anyone here used; pretty amazing. The display, okay, it's only 120x92 pixels, black and white, but it's almost like a CGA card was in the good old days. (Laughter.)

The key point is that it's the information you're interested in. You go to the PC, pick what you care about -- stocks, sports, flight times, who you want to be able to send you messages, do you want your schedule to be easily available there -- and that's exactly what shows up.

So I want to show you real quick a video that talks about SPOT and how it might be used and then Eric Lang from the SPOT group is going to come up and give you a little demo to show you why we're so excited about how this fits in with all the other things we're doing. So let's take a look.

No video? All right, well maybe we'll have the demonstration. Do we have a demonstration? (Laughter.) Here's Eric Lang. (Applause.)

ERIC LANG: Good afternoon.

So Bill told you a little bit about Smart Personal Objects and the technology behind them. I'd like to spend a little bit of time giving you an inside look at one of the first devices that we'll be releasing later this week -- how about year. (Laughter.) We would like that.

So, first of all, I want to show you a little bit about the user interface on the device. This is the watch face. You have the time channel, so you can see I can scroll up and down through the time channel, selecting custom watch faces.

One of the things I want to emphasize is that Microsoft is providing the technology behind these watches, but our watch partners are providing the actual industrial design, in fact, the style.

So, an example of why that's a good thing, let me show you this one real quick. This device is probably reminiscent of about a 1971 TI watch. It is something that we created internally just to test the software on the device, but fortunately our watch partners are creating many interesting new industrial designs for the devices, so these things will actually look good on your wrist.

So, not only do we give them the ability to design the outside of the watches, but we also give them the ability to design the inside of the watches through these custom watch faces. These are some of the watch faces that we've created internally, but the watch partners will be able to create new designs as well.

So you can see through the high-resolution display or high-ish resolution display -- (laughter) -- we're able to create some fairly interesting designs.

This one is one of my favorites. This one is called Psych Dance. Every minute it folds and unfolds. (Laughter.)

Here's another one of my favorites. This one is called Magic Eight Ball. It's also a great example of why we call them Smart Personal Objects. So let me show you how it works. Magic Eight Ball, will this demo go well? (Laughter.) Oh well.

So it's not only a nice looking watch, but it also has all the functions that you'd expect on a watch: so, for instance, chronograph, alarm, countdown timer.

But more interestingly, probably the coolest thing about these watches, is that they're connected. As Bill mentioned, we partnered with National Semiconductor to create the custom two-chip chipset for these watches, one of which is an FM sub-carrier radio. This allows us to roam anywhere across North America and constantly receive information. So anywhere I go I've got secure encrypted information coming into the watch, always fresh, always updated, always ready to look at. I never have to spend any time downloading information, and it's always ready.

A couple of examples of why this is interesting. Let's say that I go to New York. I go to New York, I get off the plane, automatically it hooks up with the time signal in New York, updates the time and the time zone for me. I don't have to do a thing. Any local information I've set up here in Seattle -- so, for instance I might have set up local news, local weather, local traffic -- will transfer with me to New York, so that means that once I'm in New York I get all the local traffic, all the local news, all the local weather, even my city guide information so I'm ready to go in New York. Everything I need to survive there is on the watch. Okay, maybe a nice weapon would be good once in a while. (Laughter.)

So, a few examples of some of the channels that I'm able to download onto the device. Here, for instance, is the messages channel. So, when a new message comes in, I'll get a pop-up alert that will tell me a message has just arrived, or I can scroll through the messages that are on there. You see here I got a flight alert. Here is a package alert. Here is a message from my friend Joe. You see the title and I can also drill into it to look at the content.

So messages, news. So, as news comes in throughout the day. I'll get updates, and again only for the categories that I've selected. So here we've got business news, national news and international news. These are just the headlines that will scroll by. And then if I want to I can dive into them to look at the actual content.

The same thing with weather. So here it's showing the current weather in Seattle. I can also get other bits of weather information. Here's the high and low temperature, three-day outlook and pressure, UV index. I can also set up some cities that I want to specifically monitor, in this case Chicago, Seattle, Atlanta, Kansas City. And it's very important for me because I travel a lot and I can get access to national weather and international weather, roughly 400 cities across the world.

So there will be other channels available as well so here are some examples of those; sports, stocks, appointments -- my appointments can come down to there.

One of the most interesting to me is the Glance Channel. The Glance Channel is a very special channel that we created to aggregate the information from all the other channels. So all the other channels will pick the most important information from that channel and serve it up to the Glance Channel.

Now, the interesting thing and why we call this the Glance Channel is that it gets the single-most important screen and then it constantly flows by and changes. So I don't have to interact with the watch at all, highly glanceable. I just look at it, I get the information or I don't. Now, if I wanted to stop it and look around I can, so here I can scroll up and down and look at some of the most important things that have been offered or I can also dive into them if I want more information. So this to me is a great channel.

So, I mentioned earlier that you have the ability to personalize these watches and the channel information that comes down to them. We do that through a Web site and I wanted to show you an example of that right now.

So, this is actually a beta site that we set up for our beta users. We have some of these watches in beta right now. And this allows our beta customers to change the information that comes down to the watch. So this is kind of an example of how the final Web site will look. Of course, it will be much nicer with a lot more channels. But it gives you an idea for where we're going. It's very difficult to customize or to set up a bunch of settings on the watch because you don't have a lot of display area there and you don't have a lot of I guess nice button interaction. So what we allow you to do on the Web site is to pick exactly what comes down to the watch. So, for instance, the watch face channel allows you to receive every so often a new downloadable watch face. The messages channel allows you to configure your messages. The news allows you to set up exactly which news feeds that you want to look at. And weather allows you to customize the weather.

Here's the news channel. We can select exactly which categories of news we want and also the content feeds within that channel. And here in weather you can obviously select your local region and some other bits of weather information. We will obviously set up My Cities so these are the cities that you want to track and whether or not you want to have support for national and international weather.

There are also other channels that we provide here.

The other interesting thing about this device is that we support downloadable channels, which means that over time we'll be extending the functionality and the channels that are available. You'll come out here to the Web site. You'll find new channels available. You click them, configure them and you can have them sent directly down to your watch.

So now I'd like to give you an idea of how we actually put all this together, a little look under the hood of the system.

One of the things that we've done is created a very small footprint of the Common Language Runtime. We did this because we don't have a lot of space on these devices. There's about 284k of RAM and about 512k of ROM so we had to have something small enough and fast enough that would allow us a lot of power on these devices, but also the flexibility to be able to do things like updating the software on the fly and taking care of security and switching between different channels and having a very nice user interface experience.

I spent a little time yesterday creating a custom watch face for this event and I'd like to show you kind of how that works. So I call this the Bill watch face and let me explain a little bit about it and then I'll show it to you. So basically it inherits from the watch team base class, and what I do here is I set up four bitmaps and kind of switch between them for kind of a crude animation.

Now, on three of the bitmaps, I create a multi-line label right here and I use the emphasize font to show that, and then I use another little helper class called flip view, which flips between different views. So I have the multi-line label in one and I have a time label in the other and the time label uses impact digits and displays in this region over here.

So we've created lots of these little helper classes to help you create software for this device, in this case a watch face.

Finally, here's a picture box view and I set it to 'no' initially and I add that to the main view.

Now then, finally I set up a timer that sets the flip timing between the different faces, the different background bitmaps.

So here it calls, the timer has a callback function, which is called face update, which is right here and it kicks off every one and a half seconds.

And then face update says if the current bitmap is zero, then select one of the other three bitmaps, grab some text from a string tag structure, and then if it's not zero then set it to zero. So basically it flips between bitmap zero and one of the other three.

And then finally we grab the correct bitmap and then we activate that view.

So here's some of the things that I've added for this particular watch face. Actually there's one missing so let me add one more.

Okay, there we go. So that's the sum total of this watch face, all the code I need. Most of it, of course, is handled in the helper functions.

We'll go ahead and rebuild this. Great.

Now there's one more thing that I need to do here. We create a special version of the portable executable that runs on this device. It actually compresses the DAT file that contains the CLR code and also blends in the resources, the fonts and the images. This allows us to seamlessly run the code in these platforms. One of the things we've done in the CLR code is that we've created it in kind of a layered fashion so it's very easy for us to port it to different platforms. In this case we changed the drivers to allow it to run as an emulator on the PC and this is where we do most of our development or we can actually flash it onto the device itself.

So let me kick off the flash. This is a little utility we wrote to actually stick the CLR code that we just compiled onto the device. I'll select the one we just built and I just need to turn the device on and off and we're flashing.

Great. So while that's cooking I'm going to show you one more thing. I'll show you the code that we just built running an emulator. There you go and there it is. (Laughter.) My own Smart Personal Bill.

So the point is that we use a version of the CLR and because of that we're able to leverage the full .NET tool chain in creating applications for these devices. It makes it very easy for us to develop code for this set of devices. We've created a very small version of CLR to run on this device. At the minimum it takes about 132k in the smallest instance and executes at about 400,000 instructions per second, so it's a very quick little engine.

Okay, so as this finishes up now we'll show it to you on the device. Okay, there we go. Now you see it in all of its high res glory. (Laughter.)

Okay and with that I'd like to ask the real Bill to come back out on stage. (Laughter, applause.)

BILL GATES: Thanks, Eric.

That's the first time we've shown actually developing a watch application and downloading it to the real device so that's great to see that operating.

There are some other ambitious technology things that we're putting a lot into. One that you've heard us talk about over the years is called Unified Storage, and this is the idea of taking the file system, the mail store, the Web page store, the directory store, all of those things and being able to put them into a single place with a rich user interface.

The degree to which this simplifies the system -- that the user can just copy one set of information, know how to navigate and search one set of information -- is quite dramatic because today whether it's a registry or the printer list or the mail, the files, it's all pretty opaque; you don't have rich tools for navigating those things. And even if you, say, want to take one machine and duplicate its state on another machine you have to take different actions in all of these different state areas.

By schematizing that state, by using the XML-ized version of SQL to do this, we take the file system to a whole new level, and so that's one of the profound advances that will come in the next generation of Windows.

There are some breakthroughs in terms of how we do testing, how we prove code correctness, how when we change code, how we do the testing processes around it. Those are critical advances for the trustworthy goal.

There is some very exciting stuff in real-time communications. Today, when people think real-time, they think calling someone on the phone. But why when you call somebody on the phone, if I have a screen and you have a screen, can't that be automatically set up so we can just work together? The majority of phone calls involve some type of list or document or map, a schedule, something that looking at that together, working on it together would be far more effective than just a voice call. Imagine talking about a bill where you don't understand your bill, something is in dispute; well, it would be a lot better to have it out there and have software on both sides helping to drive that dialogue.

So the relationship of the phone and the PC and the whole way we think about phone calls and real time communication will change. In fact, there will be a synthesis of what we think of as instant messaging today and phone calls so that the entire real time space supersets both of those things.

We're very active in that area. We recently acquired a company called Placeware to drive those things forward, so no shortage of new things to drive these digital scenarios.

So the general view here is that we see this technology applying not just in the office but also at home. We have that dual approach. We see it applying not just on the client but also on the server. We see the technology not just applying to empowerment but also reducing the cost of the infrastructure itself using Web services. We see it applying not just to the PC, but all the rich devices that are software-driven devices. And we think the basic approach that we kicked off with the PC, where you've got many hardware providers, you have a virtual layer, a platform layer and then app providers and service providers on top of that, that kind of environment that created a very high volume, low cost framework that led to the creation of the real software industry as we know it, that basic approach carries forward as we tackle all of these new things.

So, I hope you get a sense there that there will be a lot of neat things to learn, a lot of neat things to help user communities out, a lot of neat things to help guide us in how we take them forward, and we look forward to tackling these things together.

Thank you.

(Applause.)