Steve Ballmer Speech Transcript – ACM1

MR. BALLMER: It’s my great honor and privilege to have this opportunity to talk with you today. In some senses these ACM conferences are unusual in the computing industry in that they only take place every four years. And I think we all can agree that in this business four years is the closest thing to a lifetime that we could ever anticipate. There may only be one Olympics, there may only be one president, but it’s at least one major computer revolution between points in time.

I want to talk a little bit about some of the changes, I want to talk about some of the things that we see happening that will form the basis, really, for what I think of as the next revolution to hit the computing field, and that in that context talk a little bit about research, what we’re doing, what we see universities doing, and the great partnerships that I expect to see come out of that.

If you really take a look over this last four years, they don’t trust me they didn’t give me a mouse, just something that turns a light on for somebody else who has a mouse. There we go. It’s interesting, in the technology revolution they used to give me a mouse, now they give me a clicker that turns on a light for somebody they trust to click that mouse button.

If you take a look at the incredible uptake of technologies over the last four years, it’s really pretty amazing. Over 60 percent of households in the United States now have a PC, and while I think it is well worth the appropriate concern that the industry is giving to the digital divide, it is really amazing to see the rapid uptake of computers in households throughout our society. Almost half of the households that have a PC actually have more than one PC. And so the prevalence is really, really quite amazing. If you take a look at it from the standpoint of the individual, the child living in one of these households or going to school, 80 percent of younger children today will say that the PCs are critical to their lifestyle. And if you take a look at access from home plus access from school, we’re getting to the point where about 70-75 percent of all children have access to a PC. More than a third of the households in the country are on the Internet. And I think everybody sees great promise in the ability to let people access the Internet not only from their desk, but from their cell phone, and from other places as they move.

We certainly think that there will be a variety of devices that people use that access information on the Internet over the next several years. The PC will be very important. We talk about a next generation of PC, the Tablet PC, we see people scribbling on pads of paper out there in the audience. We certainly dream about the day where everybody brings a little tablet that is really a PC. We are all part of a wireless network in this room. The slides are all beamed out. The machine that picks up my voice, you can automatically just put it on top of the slides, if you want it, with your notes, your annotations. If a friend of yours is here in the room and you guys want to talk during the middle of the meeting, you can go ahead and do that. The instant messaging community will automatically be formed. There’s so much more to do.

But the advances over the course of the last four years have been truly, truly unbelievable. And never, would I say, has there been a better or more exciting time to be doing research in, development in, or otherwise participating in the computer and information technology revolution. It is absolutely exciting.

I’ve been at Microsoft 20 years, as Bob said, and we had a chance to meet then, and in some senses I think in the time I’ve been at Microsoft, we’ve had a chance to live through three important revolutions in computing, and we’re about to have a chance to enter a fourth. The personal computer really did represent a revolution in the way people work with information. And with or without Microsoft, with or without any other company, that revolution was destined to happen. In the mid-’80s/late-’80s, the graphical user interface really was a revolution. Some of that came from Apple, some of it came from Xerox, some of that came from Microsoft. The move of computing to a much more intuitive model of interaction was, in fact, inevitable, and that revolution took place.

The Internet, of course, has been around in a variety of forms for a lot of years, but it was really the move to integrate graphical user interface in the Internet through HTML in the browser led to another revolution in computing. And as we sit here in the year 2001, we would tell you that the prospects of what people will do to integrate people, Web sites, applications using XML is the next revolution on the horizon, and it is a revolution that is every bit as big, and every bit as inevitable as the PC, the Internet, or graphical user interface. And in some senses the XML revolution will be phase two of the Internet revolution much the way graphical user interface was phase two of the PC revolution.

And so as we think about what we’re going to do as a company, and where we’re going to invest, and the things that are important for us to do research on, we think about things which are enabled by or which will enable the XML revolution. And some of the paths you take to get from that statement to various activities inside Microsoft may seem a little circuitous, but they’re very important. XML will be the lingua franca of the Internet, if you will. It’s a simple format, it’s very broadly adopted. There are open standards. Some people say — Microsoft, open standards, do those things go together? The answer there is yes, absolutely yes, that’s not to say that we and other companies won’t continue to do our own patentable intellectual property. But the whole gist of the importance of XML relates to the way that things can talk together. And so it’s important that there be a set of open standards around that.

And, in fact, whether it’s the original work on XML, or some work that we’ve done in conjunction with IBM around SOAP, which is a protocol for delivering XML payloads around the Internet, we think this is a very important area for us to participate in the standards process. The hope is, you get a world where whether it’s client to server, server to server, or client to client, you have a very rich environment for building integrated, sensible, distributed applications, where we use the power of XML to do personalized presentations by user and by device, because there will be other intelligent devices, the TV and the phone, that participate in this world. And we think it really impacts the evolution of all software.

Later on, I will invite Eric Horvitz to join me on stage. Eric runs our group that does adaptive systems and interaction in research. And in a way Eric is working on the same technology base since he joined Microsoft eight years or so ago. But as you’ll see when he’s up here, the degree to which some of his research ideas and best innovations have been impacted and enhanced and extended by the potential to build off of a common XML infrastructure is really amazing.

So, I want to highlight this revolution, because I think it’s important to everybody in the computing field. People ask me, isn’t this just another thing like Java, or this, or that, or the other thing. I think this is a much bigger deal. The notion of having a standard approach to applications, Web site extensibility and integration is as big a deal as we’ll find. The fundamental advance in progress in the computing industry has come largely because people have been able to leverage and build on the work of others, and XML is the key standard that we need to have in place so that everybody’s work, if you will, can leverage and build upon the work other people do in development, in research, et cetera.

We’ve announced our platform for the XML revolution, the .NET platform. People ask me
“what is .NET?”
, and I’m not going to try to give you a detailed explanation today. Suffice it to say, it is a set of software that helps instantiate an XML platform, software that will be embedded in Windows clients, that can be embedded in non-Windows clients, we’re in discussions with third parties about that, will be embedded in Windows servers, and can live out in the Internet cloud, and help provide a framework for Web sites and clients to integrate together. How does my machine find your machine and start an XML conversation? We want an infrastructure that supports that. How do I as a user log-in and get validated by multiple Web sites in one fell swoop. The world of the future is a world in which everybody wants to bombard me with information, XML messages, how do I filter those, how do I screen out the important from the unimportant? There’s a set of software, platform software, application software, devices, that are important. And so we’re building a set of technologies which are consistent with this vision of XML as the lingua franca, and which can run client, server, service, system and application to try to make that happen.

Now, any big revolution like this starts with the notion of partnership. I talked about the fundamental importance of XML as allowing companies to build upon the work of other companies, research institutions, et cetera. And so the fundamental nature of partnership is built in. I want to talk a little bit about the kind of partnership that we’re trying to do with the academic community particularly to get started. What shall I say — the energy and sort of momentum in the computer field has always come out of academic institutions. And the need for any company to have the right kind of relationships with academia has never been higher. That is on the teaching front, it’s on the whole ways in which universities are configured, to be showcases of how technology can be used. It’s in the ways that people work together on research products, whether it’s source licensing, or technical centers, to try to support common research. And it gets down to the way companies work with other companies. I think we are pretty unique, for example, in our research environment that we allow our researchers to publish their work at any time. They are Microsoft employees, their success rises and falls with the success of the company, but there is no oversight, business oversight, on top of what our researchers, the decisions our own researchers make, on what to publish and when to publish important results in their research. And so we are actually able to have much more active research department out working with and partnering with academic institutions.

Today we have over 600 employees in research distributed over 40 different areas and in research labs in four different locations around the world. We started Microsoft Research about 10 years ago, and it has been an absolutely driving force, both in terms of our connection and communication with academia, but also as a place to incubate and start new projects and new efforts inside Microsoft. We will invest, this year, close to a quarter of a billion in research. And particularly in this tough economic climate when you see many organizations having to retrench and pull back, we’ve never felt more excited about the investments that we are making in the research area, and the kind of partnerships that our researchers have with their collaborators in academia around the world.

About 15 percent, actually a little more than 15 percent of our research budget goes directly to universities. So we’re not just talking about funding our own people doing our own research, but we try to be fairly active in funding a range of research projects going on inside academia as well as a range of fellowship programs, et cetera, for graduate students.

I was in Beijing at the end of last year, and I went with a group of folks from Microsoft Research in Beijing to visit Tsinghua University. And I’ve given speeches before in many parts of the world, and this was the first speech I’d given completely sponsored by Microsoft Research in Beijing, and I didn’t know what kind of audience they would get. They said, well, you know, we do pretty well. We’ve got a lot of people interested. The best post doc in China is being an intern here. Really, we’ll get some people to turn out. They had over 2,000 people packed into an auditorium smaller than this one on a very hot day out in this kind of suburb of Beijing where Tsinghua is based, primarily because our researchers there have been so active and involved in the community in interacting with and fostering work going on in academic institutions like Tsinghua in China.

If you take a look at the areas in which we’re investing in Microsoft Research, I think they break down into a few areas. There are a couple that aren’t on the slide, and I just want to highlight those. One I’ll come back to and one I won’t. We actually have a reasonable size group at this stage doing theory work. I can’t tell you I’m the expert on what we’re doing in the theory area. But, man, oh, man, Bill Gates and Rick Rashid can assure me it’s one of the best investments we’ll ever make.

The second area of investment for us is in programmer productivity. I’ll talk about that in a minute. And then the third area, I would say, is really in doing core research to help break down some of the fundamental barriers that exist for more rapid adoption of information technology. We talk about the reality barrier, how do you use graphics and vision to make the experience of using the computer more real world, if you will? How do we break down the barriers between people? We’re going to show you some of the research later today coming out of our collaboration and communications area. The barriers between people and computers, whether that’s natural language or speech recognition, vision, et cetera. The barriers between people and information. Most companies you visit, the number one problem people will talk to you about is, we still can’t get access to the information that they need to do their jobs, despite having all this huge investment in computers, it’s just too hard to find. Barriers between computers, one of the innovations that we’re plumbing into this version of the Windows operating system is our new very high performance TCP/IP V6 stack that was all done inside Microsoft Research.

And lastly, barriers between people at work, in the car, at home, and we’ll show you some of the work that we’re doing in adaptive systems so that my information is really my information no matter where we’re working on it.

To help me share with you some of the things we’ve got going on in Microsoft Research, I want to invite two of our researchers to join me here on stage Eric Horvitz, who runs our adaptive systems and interaction group, and Ross Cutler from our collaboration and multimedia group. I’m going to let Eric walk you through some of the work that he’s doing.

I think the kind of work we’ll show you today is some of the work that I would say we see so much more potential in, in some senses, because of the move to XML and common infrastructure and the Internet, but I know that I’m very excited about the work Eric is doing.

Eric.

MR. HORVITZ: Thanks, Steve.

I can’t remember a more exciting time for research than now given what’s going on with connectivity, with the XML revolution, and so on. And without a doubt, as many speakers today have pointed out, we’re definitely all facing this interesting and challenging world where we’re swimming in a sea of information, facing information overload.

So at Microsoft research we’ve been focusing a great deal of effort on new technologies to really give users greater control of information and communications so that they can get the right information, at the right time on the right device, formatted in just the right way. So just to give you a sense for some of the projects, here is the client of an application called Priorities that’s been under development and testing for several years now. Priorities actually learns from users how to guide alerting about email appointments and tasks. The system actually comes to understand how a user views urgency, and the systems personalized urgency analyzer looks at every email message at the back end, looks at the header structure, the body and the content, even considers the relationship between the sender and the recipient, using the Exchange Server’s org chart, to actually assign an urgency score between zero and 100 on messages.

What you see here is some recent mail that’s come into me, actually in the last couple of days, and you see the 95 here from Jenny Davidson is sort of an urgent request. She thought I could do an interview today, and I had to be here so I answered her right away, sorry for my delayed review of her email, and we’ll do the interview tomorrow. And going down here you see with lower and lower scores here we get into some

—

my Southwest Airlines confirmation from yesterday

—

MR. BALLMER: So this is a systems rating of the urgency of each of these messages.

MR. HORVITZ: Now, towards the bottom here it’s interesting we see some general bulk mail announcements about some interesting events that I might find interesting, 15 and 11. It goes lower and lower, a Stanford University presentation from an alias that I track. But, down here at 5, again, it’s a 1-800-Flowers. And let’s see over here, tired of having bad credit. One of the interesting features about having a priority filter is that it automatically is sort of a junk mail filter at the same time. So I’ve come to depend on this system. Priorities actually considers activity on desktop, as well as non-activity, as well as the acoustics, ambient acoustics in a room, and my calendar, and of course the urgency score, to decide when and where to bother me with interesting information that I might find a useful disruption, rather than suppressing the information and letting me get on with my tasks.

So the system will decide when to send email to a cell phone, for example. So when I land at Dulles Airport, for example, flip on my cell phone, and I immediately see the most urgent message from the day, I can make my calls right away. Now, I’m excited to say that some of priorities technology is available, became available publicly two weeks ago as the Microsoft Outlook Mobile Manager. And what Mobile Manager does is it lets you set up a set of profiles, it uses Priorities urgency classifiers, you set up a work profile, a home profile, an out of office profile. And the idea is you schedule that profile and then for each profile you can decide how little of your mail you want to get, only the most urgent mail, or a little bit more of that mail. And, again, I’ve really come to rely on the ability of this system. It’s become quite an intuitive secretary.

Now, if you think about it, Outlook information, email, appointments, tasks, it’s very critical in life. But, there’s more to life than just email. We have instant messaging these days, the old fashioned telephone is still relied upon, and will forever basically be relied upon for communication and collaboration. We’re interested in stocks and news service. We do some bug tracking at Microsoft on projects with a rate service. There are even other kinds of services that do agent kind of work, background querying and so on. And with the notification platform project, we’ve been exploring general architectures that let users check in with standard XML based interfaces, multiple sources of information to a generic notification manager, and also check in all the devices through device interfaces or schema.

Now, the key idea here is that to make this application smart it has to really know what’s going on in your life, your calendar, desktop activity, ambient acoustics, your location when that’s available, even a video analysis, and accellerometer analysis someday. Now, building this architecture was very exciting. It really made us think very deeply about what has to go down this wire, what XML based information should be going down this wire both into a notification manager, and out to devices. And I thought I would just briefly give you a sense for how that all works.

MR. BALLMER: I just want to emphasize, when we think about the future of information, and what we’re trying to do with XML, one of the keys is, in fact, having more information come to the user from more sources. And sort of the premium then on the kind of work that Eric has been doing just rises exponentially. This morning we did an announcement with eBay, eBay is going to integrate their Web site, which is one that people understand, very tightly with .NET. They want to be able to bombard you with notifications, your bid on eBay has been bypassed, the bids are over $200. But, then you want the set of tools that help you filter the notifications, whether they come to you from eBay, or from your broker, or whomever else in life. So I think it’s an important thing where the world says, we want more information, but then we want the rich tools to filter, which is an area where, essentially, Eric has been doing research for quite a few years now.

MR. HORVITZ: Right. I mean, if you think about it, you’d be much more willing to check in and get value out of a whole bunch of willing to help you information sources if you really had control about if, when, and how those sources would come to you, and on which device. So you see here is a little context system. You see right now the system knows that I’m facing front, my head is present, it knows that I’m here and present, if I had Outlook up right now it would see me in Outlook, if I go to PowerPoint right now it sees me in PowerPoint. It really can feel what applications are being used, and it can see me, it hears conversations going on right now. So it’s really building a rich set of contexts, it knows I’m at a desktop system. It’s looking at my calendar and so on.

Now, let’s go a little bit deeper and drill down just a bit. So first of all, the whole time people have been walking in front of this camera today, maybe Bob Metcalfe and others, this system has been tracking, looking for heads in the environment. It’s all running on a laptop here, since they all come with our vision systems lately. And it can really tell when I’m facing forward versus not being in the scene or looking away from the computer. And that’s a very useful construct to know when and how to bother me with information.

Going to the back end here now we have a context palette here that knows what device is now active. It’s looking at my calendar and sees that we’re having a meeting in my office, my spacious office that it is, and there’s activity here, multiple apps. Right now I’m at shell level, here’s a voice trace, it sees me gazing at the display. Now, that information and that context palette up here is being used to compute probabilities of different user states. Am I in a high focus solo state, a low focus solo state, are we having a conversation in my office, am I driving a car right now, is it private personal time, am I sleeping?

This is just to give you a sense for some of the abstractions we’re using to reason about if, how and when to bother a user with information, or we should say inform a user, not just bother necessarily. The system also is computing my location down over here, and with a side, you know, sort of whether I’m at my desktop or away, given what I’m seeing right now, it’s pegged out at desktop CPU. So based on that imagine we have a stream of information coming into a universal inbox or a notification inbox here from multiple sources. The idea is to stream information from multiple places, including Messenger, email, a financial source. And what we’re doing here is literally we’re pricing

—

we’re doing a cost-benefit analysis on every message coming into my cloud. And before I’m told about it the system is reasoning about when and how to send that message to me.

Now, I’m at a desktop now. The system knows that I might want to see this messenger message, coming in from Ed Cottrell in my group who has found a new problem with some module we’re building in this notification platform infrastructure. And by hovering I can tell Ed that I’ve now received his message. It’s sort of like voice mail for IMs. We let that fade away. But, in general we’re seeing financial information mixed with messenger, mixed with email, mixed with all sorts of speculative computation from agents trying to help me do my job better. Now, I should say that we’re now working closely with product teams at Microsoft. It’s a very exciting time for tech transfer on infrastructure, architecture, interfaces, XML based schema, to make this all happen on a grander scale beyond my laptop.

So with that, I think rumor has it we have a meeting coming up to the left over here with Ross Cutler from Microsoft research.

MR. BALLMER: Great, let’s go have our meeting. Ross, please join us.

MR. BALLMER: Welcome. Doug, were you going to join us, we’re all set for our meeting, I guess.

MR. CUTLER: I heard you guys were going to be talking about the ring camera, I figured I’d join you.

MR. CUTLER: So one of the things at Microsoft research that we’re developing is a new technology to enhance the way that people communicate, this involves synchronous technology such as video conferencing, but also asynchronous technologies such as meeting reporting.

MR. BALLMER: Come on, is video conferencing really that bad these days? I mean, what’s wrong with it? I admit I can’t always get through instantly, but is it that bad, Ross?

MR. CUTLER: Well, I think there is a lot of things wrong with today’s current technology in video conferencing. For instance, there’s low resolution, there’s limited fields of view, there’s latency problems and synchronization problems. So as one approach to solving some of these problems in the center of this table we have a new type of camera configuration, which we call the ring camera. This camera consists of five inexpensive 1394 Firewire cameras. They’re arranged in a pentagon fashion here so that we can get 360 degree viewpoint. So if we put this in the center of a meeting table we can capture everyone at the meeting table. And if you look in the screen up here at the top, this is the 360 degree panorama which we’re reconstructing.

MR. HORVITZ: So, Ross, what are these

—

this array of microphones over here I’m looking at?

MR. CUTLER: Also, the microphones here are primarily used for recording the people at the meeting, but we’re also using them to do sound source vocalization. That is, we can detect who is speaking, and in fact, if you’ve noticed, this window right here, this is a speaker window, it’s automatically selecting who is actually speaking within this meeting.

MR. HORVITZ: So this is all pretty impressive stuff, Ross. Is it expensive?

MR. CUTLER: No, actually it’s quite cheap. This is built with all off the shelf products. You could go and grab these 1394 boards, they’re now on the market. In fact, I built this prototype with less than $300 worth of products.

MR. HORVITZ: How about integration, there’s a lot of existing software out there today that does conferencing, does it integrate?

MR. CUTLER: This was actually developed using DirectX, and DirectShow, so in fact if you plug this camera into a Windows computer it looks just like a regular camera, except with a really high resolution we’re getting 3000 by 480 pixels, and it has a very wide aspect ratio.

MR. BALLMER: What about bandwidth, though, doesn’t this take a lot of network bandwidth?

MR. CUTLER: Actually, the way that we’ve designed this, if you look at our user interface, this is very bandwidth efficient. In fact, if you’re using this for transmitting this meeting over the Internet or Internet2, you would send a relatively low resolution version of this panorama. In fact, we’re displaying this at 1/9th the resolution that the camera is capable of. Then you send a higher resolution version of the speaker. And right now the speaker is being automatically selected. I can switch to manual mode, and just click on somebody, whoever I wanted to see. If I had a whiteboard there I could get a higher resolution version of that.

MR. BALLMER: What about other domains, besides meetings, could you use this stuff in schools? What would work?

MR. CUTLER: Sure, and in fact we’ve also taken some of this technology and applied it to the lecture room environment, so that at Microsoft Research we have a lot of lectures, both within Microsoft researchers, but also visiting researchers. So we’ve automated the capture of all the lectures now at Microsoft research. This replaces having professional dedicated people managing all the different cameras. In fact, the technology has gotten down to the point that people can’t tell whether a lecture has been recorded with our automated system or by a person.

MR. BALLMER: What can I do after I record the lectures, is there anything I can do with it?

MR. CUTLER: Yes, in fact, I think if you record lectures or record meetings, that’s really where some exciting research comes into play. So you can, for a meeting for example you can use time compression technology to speed up the meeting, so you can listen to it 50 percent faster, removing all the pauses, speeding it up without changing the pitch. You can also do annotations, add annotations to the meetings, which is good for asynchronous collaboration. You can do keyword searches on the voice track, so you run some speech recognition technology on that first, and search for

—

if you had a database of meetings you could find all the meetings where a keyword like Xbox was said. And you can do person indexing. For example, in this video right here I can click to, say, Eric. I thought Eric might have said something, and I don’t want to watch the whole video, so I just go and show me all the places where Eric spoke. And you can also do white board indexing, what we call spatial indexing. If I was recording a white board in the background here, or with another camera, I could click to the section of the white board, and it would jump me to the video section, where somebody wrote that text.

MR. BALLMER: Super. Sounds great.

MR. HORVITZ: Thanks, Ross.

(Applause.)

MR. BALLMER: The goal here was just to share with you some of the kinds of things that we’re doing in Microsoft Research, and some of our incredible enthusiasm about those. In point of fact, what we find is the independent investigation that goes on by the teams in Microsoft research are an incredible complement to what we’re doing in our development groups. And I just can’t even imagine Microsoft at this stage without the intense focus on research that we have today. There are a number of areas where we’ve already seen our research efforts touch our products. The new reading technologies that we’ve introduced, ClearType, our intelligent searching and replication technology, the new TCP/IP stack that I talked about in Windows 2000. The Outlook Mobile Manager that Eric had a chance to talk a little bit about.

We’re incubating, or we have incubated, some major projects. All of the Windows Media Technologies work was originally incubated almost ten years ago in Microsoft Research. Our natural language and speech stuff we’re on the verge of shipping a new release of Microsoft Office where literally if you speak Japanese, Chinese, or Korean the dictation is probably about 4 times faster for text input than what even a fairly accomplished Asian person could do in terms of text entry for those kinds of more complicated languages. Windows CE, this Tablet PC effort I talked about, those have all been incubated inside Microsoft research. So we’ve got a lot of enthusiasm for the work going on there. Some of the most important work that we’re doing has yet to see the full light of day outside Microsoft, but is literally transforming the way we build our own products.

These days, for good or for bad in some sense, our products are literally millions and millions of lines of code, and the complexity of developing and verifying and testing those products has really gone sky high. And so, if you ask us what is the key investment we need to make as an organization, and to productize for our customers and research partners, it’s in the tools to improve the basic productivity of a developer. We have built in research a specialized group that focus in on programmer productivity, and I will tell you there’s absolutely fundamental innovations which allow us to continue to even build and think about projects of the scale of those that we’re doing that could not be attempted without some of the tools that we’re doing for performance analysis, in verification, as well as in development and correctness verification, tools coming out of the Microsoft Research productivity, programmer productivity center.

I want to talk about one other effort where research is meeting some of our work in the XML and .NET world, and that’s a project around what we call the TerraServer. TerraServer is a very large spacial database that was brought up on the Web by Microsoft Research three or four years ago now, and the first thing that the group wanted to do was to really prove out that you could run very, very, very large data sets, spatial data sets on Microsoft SQL Server over the Internet. And so we, our research team, built this thing as kind of a small experimental project a few years back. And that Web site actually now is a very popular Web site for people doing certain kinds of geographic mapping, and for some science type projects on the Web. It’s the largest single database in terms of amount of information stored on the Internet.

What we have done of late, again prototyping and experimenting with this TerraServer Web site, is to expose it as a set of XML Web services over the Internet. So you literally can take now population data, for example, and overlay it on top of these photographic images that we store in TerraServer to produce customized views of this spatial geographic information that we provide now as a service to other Web sites on the Internet. There’s work that we’ve done with the U.S. Geological Survey, this is being used today in some distributed computing courses at MIT, and it really is sort of a very good example of showing off the benefits of having Web site or applications be able to integrate and leverage off of work that’s already been done.

If somebody wants to take our geographic database that we’re running, we’re operating, we’re taking care of on the Internet, the cost to include that or to add information to it from their own Web site is very low. We don’t have to build anything. They don’t have to deploy anything. They don’t have to operate anything. They just have to write in an XML Web service. And it’s all standards based. It uses XML, it uses SOAP, it uses the UDDI protocol so that people can discover it and understand how to program against it over the Web. And so, again, our research group is taking a pioneering role in trying to prove out some of these concepts.

I talked about the partnership with academia as being fundamental. And I really mean that. We’ll spend this year about 75 million in research and education. We sponsor academic conferences, we did some work with the ACM on a programming contest, for example, at MIT. We’re doing work to promote .NET as a research and a teaching platform. I’ll talk a little bit more about that in a minute. And with MIT particularly, we’ve done some work not only on, if you will, the computer science of what we do, but how do we use technology to transform the learning experience.

As Ross was doing the demonstration, I have to say, I had to bite my tongue not to make a small joke about my own academic experience. I decided after a couple of years at Harvard, there were things I liked to do better than attend actual course lectures. You give me a course lecture online recorded in the way Ross talked about, you give me the right tools to parse through it, I would have been just the holy terror, I guess, in terms of my ability to cut classes. And I don’t mean that in an offensive way. I mean it in a way that’s really designed to be thought provoking, giving people the kinds of tools that lets them sift through and parse through large amounts of information more quickly is absolutely fundamental in the academic sphere as it is in the business sphere and every other.

We do have a range of university collaboration underway. And I want to highlight some of them. CMU has a big project to do, a large scale wireless network, campus wireless network. We did work where we provided the source code of Windows CE to CMU. They put in place all of the Kerberos authentication technology, and then licensed this back to us.

With the Florida Institute of Technology, we’ve done work in virus detection and neutralization. With Cornell we’ve done a set of work around building super-computer style design around Windows 2000. With Berkeley, we’ve done a set of projects in the area of improving usability and quality of life through information technology. These are serious and important projects to us, and I know certainly to the universities with whom we do them. The commitments range from large to small. We can’t fund everything people want us to do. But it certainly is a very important part of in the overall scheme of things that we fund in the research area.

For people who have been coming to ACM meetings for a while, I think there’s a lot of good news. With the XML revolution sort of in full swing, software has never been more important than before. We’re certainly investing in that. We’re investing in that on our own, in partnership with universities and research organizations, and we’re trying to invest in a way, particularly in our research organization, where we’re publishing and sharing out ideas with you quite liberally, and we’re benefiting back from the great work and things going on in all of the institutions represented here today.

It’s been my pleasure to have a chance to make these formal remarks. I’ll look forward to the question and answer session. I know with Bob moderating, it will be certainly energetic and active, if nothing else.

Thanks very much.

(Applause.)