MR. GATES: Well, good morning. It's great to be here and to have a chance to share some of the exciting opportunities we all have ahead. I'm pleased to do this in connection with our partner NCR.

Like all of you, we're dependent on NCR for some very key things. Our worldwide network is being upgraded, and the quality of that network is completely managed by NCR, which is about as critical a thing as we have in maintaining our fast moving operations.

I'd say today, the way that technology is going to change how business is done will be more impacted in the next 10 years than it has been in the last 20. We're really on the verge of some very important changes, and companies that seize those changes in this information age will be the ones that lead.

It's not really about the hardware technology at this point, because the improvements have been so substantial, the communications speeds, the power of the PCs, the standards that connect these things together, that almost any imagination you have about how information should flow inside your company, or between you and your customers, it can be achieved today. And it's companies that step back and really think that through and use that for competitive advantage that are going to be the most successful.

This vision that we call digital nervous system encompasses all the kinds of information inside the company. Historically, we always thought of the computer systems as dealing with transactions, and then the PCs being out there to run productivity software. But the real glue between those, analyzing that data, letting people dive into it to see it at a detailed level, being able to mail around different views to one another, even letting customers come in electronically so that they don't have to get their information on paper or call in, all of that is encompassed in what I'm calling digital nervous system. So, it's the basic operations as well as the strategic thinking, as well as the customer interaction. And all of this really defines the business reflexes of the company. How well does the company deal with a surprise, with a change in business conditions, with a new product from a competitor, with a customer who is calling in and asking for things to be done in a better and different way.

So, there's many principles, I think, that are key as companies are designing these systems. These systems have to get all the news around very quickly so that people are coordinated in what they're doing. Now, getting good news around is very simple. I get mail all the time from people saying, hey, isn't it great, we've just finished a product, or we just won an account. Whenever I get those messages, I say to myself, well, that's interesting, they only sent mail about two or three accounts we won today. Does that mean that we lost all the other accounts? You know, what really is going on.

And if you think about it, what's actionable? The fact that things are going well, which gets around very quickly, simply reinforces that you should keep doing exactly what you're already doing. Particularly in our business that's fast changing, it's those symptoms, the bad news, where somebody expected more or another company did a better job, that really might change your plans and let you adapt to the market requirements.

So, one principle is, bad news must travel fast. And good systems can overcome sort of the human instinct that, you know, it's not as fun to get that information around.

A second principle is, the Internet changed everything. It is now possible within your company or to your customers to have all the information there in a rich, electronic form. Paper invoices, paper billing, paper forms, these things are going to go away. And within only a few years, they'll be more the exception than the rule. The idea of empowering workers is key to this. And the Internet is growing at a pretty unbelievable rate. I think there's a tendency with all new technologies to sort of expect more out of it in the next two years than it's likely to deliver, but to underestimate over a longer period, five to 10 years, what it will do.

In the case of the Internet, yes, in the next few years it will probably take more time for people to get used to electronic mail, to get comfortable with the set up, to get the kind of high-speed connections that make it just obvious that every day that's the way you go to get your information. But as the generation that has been educated in college with the Internet as a standard tool comes into the workforce, as the incredible technology improves the way we interact with these machines, as the electronic data exchange standards get to be very strong, all of that will be like a revolution. And so, in that time frame, the five to 10 year time frame, I think people are still underestimating the impact.

Another principle is, customers should be at the center. Of course, this resonates perfectly with the expertise that NCR has in customer relationship management. Getting all the information about the customer into one place, being able to look for patterns, being able to customize both the electronic interaction of that customer, and to help the human-to-human interactions be really related to the needs of that customer, or what opportunities you have to sell something additional to them, that is fundamental. In most businesses, that customer relationship will become the primary asset. And so, it's simply information that's not being used in the way that it can be and will be by the leading companies.

Finally is the idea that every worker is a knowledge worker. You should make their jobs interesting, you should draw on their ability to work with customers in a straightforward way where decisions can be made very quickly. Any jobs that were just data entry jobs, those are the jobs that go away. Once you get the Internet, there's no need for that just straightforward entry of information. And so, you can take people and broaden what they do, and treat them as people who need the information to help your customers.

What are the pieces that go into this? Well, I'm not suggesting in any way a radical overhaul or a big investment. Most companies today are already buying the building blocks, and yet there's a huge difference, certainly a factor of 10 difference, between the value that people get out of these building blocks. People put up-to-date PCs on people's desktops. They connect them together with a network. They have transaction databases. But, in many cases, all of that is used just for the basics, just for productivity software, and just for printing the bills. It's not used as a way of exchanging information in the company about project status. It's not used as a way to get that customer feedback in and really use that to define how products should be defined.

One way to look at this is to say, new companies don't have paper forms. When you want to change something related to your personnel records, when you want to check on a project status, when you want to get the sales information, is there paper involved in that? If so, you don't have the rich views. You don't have the ability to share easily with people at a distance. And so, you're falling far short of what that substantial investment can provide. Other key elements here include the line of business applications, and a data warehouse that's gathering together the information across all your product lines and letting you determine exactly how you should relate to these customers.

And so, there are many ways that we can rate a digital nervous system. It's very clear to me that even companies that are very focused on this are far short of realizing the full potential. At Microsoft, if somebody wants to do business with us, they enter their information electronically. We don't take paper bills or paper invoices. So we're getting there. When we do something like our United Way campaign, charitable giving, electronically we notify people, and we give them the ability to browse an unbelievable amount of information about the various agencies, so they can be informed about designating their gifts, or getting involved as a volunteer. So, even with something straightforward like that, the electronic approach is far better.

One metric I'd propose is that if a company makes it hard to go and find old information that no one will take advantage of that. So, if you can't within 60 seconds sit at your PC and call up a post-mortem analyzing a previous product rollout, then you're very unlikely to take the trouble to go and find it and benefit from whatever learning got done there. And so, corporate memory requires that the documents be easy to find and online.

Do managers share the same information? Are the people at the top of the company and the people who are at the operating level seeing the same data? Historically, people talked about executive information systems, but those things, by confining themselves to a small set of employees, all they did was invite confusion, where the executive would see one number and be concerned, and then the employees who he would call up and say, what are we going to do about this, would be looking at a different way of presenting that data. The data should all be there, so you can get down to the lowest detail or up to the highest summary, ad hoc, in fact, just pivoting around, grabbing that data and mailing it around to other people.

We've talked about the customer information. A key element of this, I want to emphasize, is the idea of customer feedback. If your product is popular, your customers are willing to take time to tell you how they'd like to see variations of it, how they'd like to see it change. And that is a huge advantage. And any company that takes advantage of that is certainly going to stay on top in the business they're in.

However, because of the size of organizations, it's very typical for that information to get lost. It's given to the salesperson, or it's mentioned to someone who is out in a meeting with that customer. But it doesn't get back in to the people who are actually engineering up what the next round of the product will look like. And that's an incredible waste that a digital system ought to completely avoid.

More and more there's a key requirement now to bring partners into these systems. Again, doing that digitally has got to be better than trying to do it through paperwork. In fact, the new economy requires that we use partners more creatively, that we look at lots of things that were done inside the company, and consider doing them outside with somebody who's got more of a focus on that. And yet, you need to be able to do that without giving up a complete view of what's going on, what's happening with your customers. And so the partners have to be drawn in as well.

Perhaps the greatest test of a digital nervous system is when something goes wrong, the various crises I talked about. Can a company within 24 hours change the product pricing, or change the delivery orders, in order to deal with the new situation. Well, that's going to require many people in different locations, disagreeing, looking at the data, diving down into it, and only a state of the art nervous system would really let you do that in the right way.

So what is the role of the PC industry and Microsoft in this? Well, we're providing high-volume building blocks, building blocks that are sold in the hundreds of millions, that allow us to do massive R&D, billions of dollars of R&D, and yet charge far lower prices for these things than ever has been seen before. You know, the operating systems we have today, we put more R&D into those operating systems than are put into, say, MDS, which is of course a very high end, expensive system. We put more in, but because of volume the price is 1000 times lower, and therefore you get the volume, you get the applications built on top of that. Specialization has been key to the success of the PC industry. Specialization in chips, specialization in different product areas, and it's all those specialized companies coming together that have made the PC now the mainstream in computing.

Our building blocks are very well known, Windows, Office, BackOffice, and we constantly improve those products and get applications and value added solutions built on top of them by partners. We ourselves are not in most of these businesses. We're not in the vertical businesses. We're not in the enterprise application businesses. But, our partners are, people like NCR have expertise across a number of areas where they're taking advantage of the work we do.

Well, we have initiatives that span all of these products. Not only do we come up with the new versions of the individual ones, but there are things that really drive what goes into those new releases. During 1996 and 1997, a big priority for us was the Internet. It's still a very important priority, but during that period it was a jihad where we have to really build Internet standards and get ahead for all the products. That was very, very successful. And we continue that, in fact, we're more involved in those standards committees than any other company. And there's a lot that needs to be done, whether it's security, quality of service. So the Internet continues to evolve. Today some of the priorities also include simplicity, reliability, scalability and interoperability. Those are, together with the Internet, the top five, and I want to touch on each of them.

The top of the list today is simplicity. Delivering the increasing power of these machines, without forcing people to learn more commands, more utilities, just more difficult situations that they have to understand. If you look at the complexity of installing, updating, connecting, dealing with the case of buying a new machine, you're really being exposed to a lot of things that shouldn't be necessary. An extreme example of this is to go through some of the error messages that the system can deliver.

The one I'm showing here is my favorite, it says, the DHCP client could not obtain an IP address. If you want to see DHCP messages in the future choose yes, otherwise choose no. Well, the message is very helpful if you're confused about yes and no. But, if you don't know what DHCP is, which people shouldn't have to know, it's completely cryptic. Even to somebody who understands DHCP, you're confused. Is something going wrong here, and what does it mean if you ask to not see those messages?

So across the board, we have to take the scenarios and reduce rather dramatically the number of commands that people have to deal with. Now, this priority got set about 18 months ago. And so the round of products that are coming out in the next year, the new version of SQL, the new version of Office, the new version of NT, do make substantial progress. Things like a directory to manage application deployment, or a mirroring technology that makes sure that all the data on clients is automatically transferred up to the server, so you can always replace a machine, or have the information backed up, or access the information from another person's machine, without even having to think about it, those are very, very helpful steps.

Now, I think we have a long way to go here. In fact, we've gone out and looked at what it's like when these technologies come together in somebody's home or somebody's business. And I'd like to show you a video that sort of illustrates how complex it can be today.

(Video shown.)

MR. GATES: A lot of work to do there on simplicity. The next initiative I mentioned is scalability. And here I think people really are going to be surprised in the next couple of years, both on the hardware and software side. On the hardware side, Moore's law continues at full speed with increased clock speeds, much more efficient processors, and now for the first time we'll have 64-bit processors, including Intel's Merced. Not only will the individual processors be far more powerful, but we'll be combining more of those processors together in a single system, moving up from a typical four processor system, to 16, 32, even 128 process systems. That will be done by moving away from classic BUS structures to point-to-point communications, and an approach called non-uniform memory architecture or NUMA. And that's just around the corner. Having Windows NT systems with an incredible number of processors.

Now, each of those systems can be connected together into a cluster. Today we've had two-node clusters for reliability fail-over. But, next year we'll move up to have 16 and then 64 node clusters. So you take all of those improvements and combine them together. The chip level, the number of chips, and the number of systems in a cluster, and that's the contribution the hardware side brings.

MR. GATES: (In progress.) On the software side there's constant improvement in the software algorithms that also provide additional performance. But, in order to tackle these high-end problems, there's a lot of pieces. Dynamic clustering, the diagnosis, the online backup, extreme quality in the drivers, a level of certification for these high end servers that will go way beyond what you have for typical drivers used in server systems. These have got to be systems that run for years without reboot, and that's what we're very focused on in the next round of NT.

As I mentioned, the R&D we have on NT is massively greater than what's been applied to high end UNIX or any other system. And so its scalability, I think people are underestimating what's possible. You can have the PC architecture extend from the very low end up even beyond what the most expensive systems have been able to do. And the benefits to that in terms of applications development, common user interface, common approach to all of your systems will be very, very substantial.

Now, in this world where Windows is taking on a major role, of course there are going to continue to be lots of other systems. And part of the utility of Windows is how well it interacts with those systems. We've had a lot of initiatives every year to strengthen those interconnections. And several of those that I'll talk about are done together with NCR, drawing on their expertise, because they have the experience of working with all these different system types. We do support all the different databases with our tools, not just our own but all the popular databases, now we have built-in driver support.

We support UNIX systems being connected up through the directory. We support transaction gatewaying into CICS or any of those other transaction systems. We support the object model, the COM model on all the different systems. We bridge that into CORBA. And we also have initiatives in each vertical area to take and to find a set of objects so that information can be shared between applications. Active store is a good example of that in the retail area. And NCR is really our lead partner in terms of helping us to find those things to meet special things customers are doing, particularly what are called their value chain initiatives.

A final interoperability effort is to take XML and use it to describe the meta data that will allow you to take data from your transaction database and move it over to your data warehouse database very easily. So there's no manual intervention required there whatsoever. XML is a tool that will make that far easier than it's been in the past, and we're pleased to be working in partnership with NCR to define some of those standards.

The last area I talked about is the manageability. There's a lot here because people in the past have had to go out and visit individual PCs; individual PCs have been very different from one another. Getting even just a patch out to systems has been hard. Being able to monitor what's going on, what's not working, has been fairly complex. Well, that's a software problem, and it's one that we've made some huge progress on with what we call the management console. This is another place where, as Teradata is now available out on NT, they're using the management console, so they unify together in one place how you can handle your Teradata information with the other information that's in the system.

I've got a little bit of a history here of the different things we've been doing with NCR. It really goes way back, even before 1984, the first interpretive basic I did in 1975, NCR was as customer for that, for one of their data terminal products, actually the second customer that Microsoft had. When we started out with Windows NT, which is now about six-seven years ago, NCR was there at the beginning. We've had a lot of work in vertical areas, a lot of work particularly with data warehousing. And this recent development, which is actually having Teradata shipping on NT is a great milestone in terms of the things that we're doing together.

I just picked a few examples here of customers that have solutions in place, and a few that are projects that we're doing together right now. And you'll see there's quite a span of things here. Data warehouse is present in a very high percentage of these accounts. Derendinger AG is a case where it's an SAP implementation and the data is actually moved into special NCR software for the analysis. Office Depot's big roll out, Smart and Final is another big roll out.

Exchange is a major product for us, there have been a number of customers where people are tying that in. I view the mail system as sort of just the beginning of a company wanting to have a digital nervous system, getting every employee to use mail every day really leads to them thinking about, well, what kind of information could be there. But, it's got to be reliable, it's got to be from the top of company and an example set, this is the way information will flow. The National Football League is using it, they have over 30,000 messages a day, about players' statistics and trades and new things going on. And so it's become very central to their operations.

Sprint is a great example using this Teradata technology, where they really have to respond very quickly because of the competitive offers that come along. And the setup that's being built for them will be very helpful there. Bank of America is a huge Teradata database. Nextel's customer care, K-Mart is a retail self-service. So we've really come together to bring our strengths, which are in high volume software building blocks, that are standards that gets lots and lots of applications along with the industry expertise and the ability to work very directly on particular solutions that NCR makes available.

There's one particular project that I thought would be a lot of fun to show you. And that's some work that NCR is doing on a retail self-service solution. So I'd like to ask Brian Francis who has been the lead developer on this, to come up and give us a quick look.

Hi, Brian.

MR. FRANCIS: Hi, Bill. How are you doing.

What we'd like to take a look at today is the new NCR self-service solution that we have packaged here. And there's two critical issues that I want to touch on before we look at it. One is the hardware and really the underpinnings of this solution. We've got an integrated package, with an LCD touch screen, Intel processor, basically a fit or smart client device, that's running Windows NT for an operating system, and then is tied into the back end, using the Windows DNA architecture.

So we're leveraging technologies such as dynamic HTML, and active server pages, and then using COM to allow us to integrate into data sources and to existing legacy applications. But really, more than just the tools that allow us to build this, we've taken a lot of time and effort to develop a set of solutions, and working with both the users of the solutions, and the retailers themselves, they've really come up with a very effective, and available solution for this.

So in our processes one of the things that we do is we create scenarios that identify how people will interact with devices such as this. So in this example we've created a retailer called Witson and Witson is your Best Buy or Circuit City type of electronics retailer. And this device would be running, let's say, inside the VCR department that allows you to get information about VCRs, and also see promotion.

So one of the things that our customer research showed that people are interested in finding out is they really want to know what's on sale. When they go into a department, they really want to know what's on sale. So by a simple touch of the screen, we can go and take a look at all the different VCRs that are on sale. Now, one of the other advantages of taking this type of product information and tying it into a data warehouse is that you can track the likes and the needs of a customer. Now, if you take our your Witson's card, and you swipe your Witson's card, we can query into the database and come up with targeted promotion offers for some of the different VCRs. And we take a look and get more information about this VCR. And at this point we've used COM to integrate into the existing retailer systems, to pull out things like pricing information, availability, product information, and location.

Now, one of the advantages of an interaction point like this is that we're able to track what a customer not only purchases, which has been the traditional use of a data warehouse, but actually what are the products they looked at. So you can analyze and say, well, this person looked at these seven VCRs and ended up purchasing this one, so you can identify trends as to why certain VCRs were picked, as opposed to others. In this example, we actually understand and know the products that the person has purchased in the past, and we can tell them that, if you want to purchase this VCR that you're looking at, you're going to have compatibility problems with your existing stereo system.

Now, one of the other things that this device allows us to do is really to supplement the store associates. So maybe you're interested in picking out a VCR, but you really don't want to go through the all the little technical details. So you want to touch on the -- we've got a little wizard that will come up and step you through some of the questions. So what he's going to do, he's going to ask you a couple of questions, and why don't you go ahead and respond to some of these.

MR. GATES: Okay.

MR. FRANCIS: So based on your responses, what we've done is we've taken the product information and built a set of data models. And the system will take your responses, and match those up with a set of VCRs that are available that match their information. And then based on one of the VCRs that you're interested in, again, we pull all that information out and present it to the customer.

MR. GATES: So this is actually an NT system running behind here, and you've built on top of that?

MR. FRANCIS: We have NT running on the client. The display that you see here is actually Internet Explorer. We've taken off all of the caption bars and the menu bars, so we're just giving the user a very interactive experience that they can follow. And then this ties into a Windows NT Server, which can then tie into, using COM or DCOM, the existing retailer's applications, or all the way back to the data warehouse at the corporate location.

MR. GATES: And how did the retailers respond to that?

MR. FRANCIS: They've been very enthusiastic about that. We've had one of the retailers you mentioned, K-Mart, they're an early adopter of this technology. They're very excited about it. It offers a wide range of possibilities. We've built a lot of multimedia capabilities into the device. So not only can you use this as a point of information, but in working with your vendors, you can provide them with a very targeted promotion engine, from playing high resolution video, and product information, making it available to the customers.

MR. GATES: It's great to see that. Thanks very much, Brian.

MR. FRANCIS: Thank you.

MR. GATES: (In progress.) Well, our partnership has a lot of exciting opportunities ahead. More solutions using BackOffice, getting all the different NCR value-added software onto the new version of NT, taking advantage of the extensions there, whether it's Light Keeper or Teradata or Master Minder, making it very easy for people to take transaction information from our SQL server and get it up into the Teradata data warehouse by using these XML schemas, and using the repository, and the rich replication that is now available there.

We're building 10 centers where we can actually show people examples of things we've done together, the latest of our technology and how that works. And we're really pleased with the emphasis that NCR is putting on adding value to Windows NT.

As we think about this next decade, the way that we interact with these computers is going to be very different. Computers today are defined by the keyboard and the mouse. But, with the extra processing power, and with some magic software, we'll be able to define new modes of interaction that are far more natural. Breakthroughs in screens with software will let us have tablet type devices, where we can carry those around, handwrite on the surface of the device, fill things out, take them to meetings, and so that will just be another way that will extend the digital system.

Speech recognition is finally coming into its own, both for dictation and command and control, and certainly it's just a question of when, not if. Even things like vision, where the system sees who's coming up to it, what mood are they in, what are they trying to get done, what gestures are they making. That's actually very feasible because of the incredible processor power that we have.

Microsoft spends almost 20 percent of its revenue on R&D in the different areas I'm showing here, advanced graphics, decision theory, because we are so much just scratching the surface of what it's possible for software to do.

So, the bottom line for all of you is, there is a great opportunities to lead inside your company to build these digital nervous systems, both for internal sharing and working with customers and partners. These systems will just continue to get better and better. We're certainly very committed to the enterprise scale. That's a point that we make very clear on all the new investments we're making, enterprise is very crucial to us because that's where we think these great new applications will be proven out. And we're pleased to be doing this with an excellent partner in NCR, and that's why I'm very happy to be here today.

Thank you.

(Applause.)

MR. GATES: I think time is limited, but we probably do have time to do some questions. I think we've got four mikes. I haven't seen all four yet. But if people are interested, if they want to get up and stand by the mike, I'll call the mike number and answer some questions. I think I see somebody at number one there.

QUESTION: Mr. Gates, recently you've had a change in your leadership team, and Steve Ballmer has taken on a new role. How do you see that impacting your customers, your partners, and Microsoft itself in the future?

MR. GATES: Well, Steve's stepping up to president was a great thing for me. Steve, of course, has been with the company over 15 years, throughout most of our history. And so you'd have to call this more of an evolutionary change than something radical. It is going to let me spend the majority of my time with the product groups. So looking at what is NT 6, how do we get the unification of all the different subsystems so that you don't have all these utilities, so you don't have all these commands. How quickly should we take things out of research and get them into the product? How soon do we want to bet on speech as a fundamental way of interaction? And so, Steve, in stepping up, there's a lot of great people who work for Steve, like Jeff Raikes, who's job responsibilities got increased, and so he's got the business management responsibility.

I'm still CEO, and the buck definitely stops with me on all things, but I will get to balance my time in a bit of a different way. Because of the dynamic nature of our business, we find ourselves reorganizing, to some degree, at least about once a year. You know, bringing emphasis to new customer units, getting the product groups that need to really be working closely together under common management, and, you know, sort of time to time we make those adjustments. This one was definitely a very positive one, even in the last few months I've had a lot more time to do the things I enjoy, where I think I can have a major impact.

QUESTION: You've dominated in the PC arena. How do you see Microsoft playing in the data warehousing arena in the future?

MR. GATES: Well, data warehousing is, we think, a fantastic application, one that's going to grow in popularity quite dramatically. We do see Windows NT as being a key building block there. We do see for small data warehouses, people will be able to do that with SQL server. As they move up and want to do more sophisticated things, that's not a specialized area that we're involved in, and that's why we're very glad to have NCR taking their Teradata software, putting that on Windows NT, and making that be a very focused solution, including all the services they have that go around that.

So, we are very much in the database business. We'll make sure Oracle's life is kept interesting and very competitive.

(Applause.)

MR. GATES: But when it comes to the very specialized requirements that people have in data warehouses, that's not something that we'll be going out and bringing the pieces together, and that's why the partnership model is key to us. That's really the most unusual thing about Microsoft is, you know, all the things we've decided not to do, no chips, no systems, no -- you know, we're not a big computer services organization like IBM, or even Oracle has chosen to become, we've chosen to say, hey, there are people who know those things better, and data warehousing is one where the partnership leverage is a great way for us to make sure our building blocks get used.

QUESTION: I really appreciated the slides that we saw on the video for there's a lot of remotes for the various components. In the world of industry, we have the current situation where we want to integrate -- (inaudible) -- now, we all know that the preferred tool of each and every customer in our business is Microsoft Excel, because they're comfortable with that, and they know it and they want to use it.

Now, I'm trying to show to them all these fancy tools and so on, but they always go back to Microsoft Excel. When can we expect a few additional advances with functionality on Microsoft Excel just to cater for those little bit more need for the customer, so that we can get rid of all these additional software add-ins and so on?

MR. GATES: I think that's a very good point. Excel is very, very popular as a data navigation tool, and it lets you do pivots, it let's you dive in. It understands fairly rich data navigation. Most of the data analysis tools, whether it's data warehousing or otherwise, have been built separately from Excel, and so you have to go into a new world and learn a different interface.

As we've made Excel more extensible and built in some of these richer features, people are now recognizing that they should do that user interface value-added on top of Excel. And, of course, we want to encourage that. WE want to get their feedback on what we need to build into Excel to make that a more straightforward thing for them.

In the next round of Office, which is called Office 2000, and comes out early next year, Excel takes a huge step forward in this direction. So, we'd love to see the data warehousing tools more and more integrated in on top of Excel, and I definitely think you will see quite a bit of progress in that area.

QUESTION: Mr. Gates, one more question please.

MR. GATES: Okay, one more.

QUESTION: When you talked about the bad news should travel fast, and on the other hand you talked about the digital nervous system, how do you feel this thing works in your company at this moment? Like, okay, some accounts are won, some accounts are lost. How do you get to know, how does bad news travel, and how fast are Microsoft's reflexes?

MR. GATES: Well, the electronic mail system is a huge enabler because, you know, everybody can send information around very easily. You still have to have culturally people be used to the idea that if they send out bad news that the messenger won't be shot, that people are going to take it into consideration and really understand what's going on there. We also have our account tracking system totally online, so you can go up and look at the status of every account.

I think we do very well in terms of getting the bad news around. I mean, if anything, it's our culture, people sort of emphasize the bad news. Sometimes they go overboard. Every once in awhile I have to say, okay, okay, we're actually still in business, folks.

(Laughter.)

MR. GATES: You know, it's not that bleak. In technology, you always want to be looking for the changes that are going to redefine the framework because the technology companies, they often have their best years when they're making their biggest mistakes because there's this lag between missing something when it's early, and then it having a huge impact.

You know, we know, we're painfully aware of where we need to make things simpler, where cost of ownership is too high, where we need to pull things together. And I think that's because we allow this information to flow quite well. But it's an act of daily management to make sure that the feelings of the customers, the morale within a group, the feelings about sharing technology with a company, all of those come to the surface, and people in the extreme case, you know, they're willing to send me mail if they think there's something that's off track.

So, I think so far we've done very well on it. But it's something that requires ongoing vigilance. I do think we're better at bad news traveling fast than most other companies. It's one of the few things I really would say is a best practice that we're definitely sharing with everyone.

Thank you.

(Applause and end of presentation.)