Back to agenda

Financial Analyst Meeting 2005
July 28, 2005

ED LAZOWSKA: I'm Ed Lazowska. I'm a professor of computer science at the University of Washington, and it's a great pleasure to be able to talk Bill and Ray this morning, and hopefully you'll enjoy and learn from the conversation. I'm going to start with you, Ray. You're sort of the guru of productivity software and collaboration software, and we've have Notes and Groove and “Symphony,” and I don't think of you as a big company guy. Someone pointed out to me yesterday that you started at Microsoft in the middle of April, so you've passed your hundred day mark and the honeymoon is over. So as you've spent the last four weeks out of five here in Redmond and turned over the rocks, let me know what you've discovered about Microsoft that maybe you didn't expect and what things you did expect.

RAY OZZIE: I've worked with the company for many, many years, and I didn't expect to be surprised much in terms of the technology, because we've worked both at the edge of the organization and I've had relationships with people at the top for quite some years. What I think I had really no awareness of and I've gained a tremendous amount of respect for over the last couple months, is the management processes, the decision-making processes at the top to ensure execution while at the same time making significant investment decisions to ensure that we're staying on a growth path. It's very interesting on a meeting-by-meeting-by-meeting basis, understanding and watching people balance those two aspects of decisions.

ED LAZOWSKA: So I wonder about this, because you talked to sort of smaller go-go companies, and they have a management style that, it seems to me, doesn't scale. And you've been with start-ups, with medium-sized companies, now you've looked at a very large company. Can you talk about sort of the scalability?

RAY OZZIE: It feels, at every level of the company, more like a start-up, than it does a big company. I did work for IBM for a short stretch, and while I can't really compare the two, you can see down to the individual, the passion and the motivation of wanting to continue to move things forward, move things forward, go into this area while continuing to execute.

ED LAZOWSKA: Bill, have there been changes over the past, say, five years in how the company's organized and manages itself to try and keep things going in that direction?

BILL GATES: Well, definitely the challenge of having such a broad technology portfolio, and getting the benefits of where those things merge together, you know that's something we come in and work on every day. It's very cool to see a lot of these things coming together. For example, in e-mail we started with Hotmail as our kind of low-end free thing, and Exchange was this high-end corporate thing. And we're really saying, okay, our customers should be able to have deep, rich e-mail services whether they run the servers or whether we run the servers for them. That is really the best of what Hotmail's been—where you can just connect up, create an account very easily—and what Exchange has been in terms of capabilities. And so now we want to say, okay, you can buy that as a server, you can buy that as a service. We're not fully there on that, but it's actually the excellence of Exchange as a server product and Hotmail as a service-type product. That's big-scale—hundreds of millions of users—it's really the learning from both of those coming together that will let us do something pretty neat.

And I think Xbox is another good example of this. Of course, it's the video game business, very special business; we've had to protect that brand as a brand that stands for something unique. But underneath Xbox, they way they do voice and instant messaging and video and graphics, and the way it's going to tie into the Media Center, just be the extender, automatically, if you have a Media Center — boom, you can get even high-definition there, that's the kind of integrated thing we can do.

And so, I'm very close to this, every day I'm saying to group A and B, Hey, we can get more out of the things that are going on between these two groups. Right now there's a lot of that going on between the Longhorn—Windows Vista group—and the Office 12 group, exactly making sure that the products are very, very complementary to each other.

So I do think it's something we've refined over the years, and it was a big part of the change we made in 2000, where I went to focus full-time on those issues of product strategy, and Steve, taking over running the business, said how do we have a more scaled management system, how do we draw more people in, creating the seven P&Ls to drive some responsibility down, changing how the P&Ls looked at the sales force so they could control how much sales force activity there was and make tradeoffs between marketing, sales-force, R&D, and those things.

So we've always wanted to be a company that got the benefit of the scale, going back to the beginning where we said, hey we're not a one-product company. Many of our competitors were one-product companies. Now they're not one-product companies, but Sony's just in video games, Nokia's largely just in mobile phones, IBM's largely in enterprise type software situations. And so is it a benefit to us that we work across these realms that we can expose you to the same user interface at home that you get at work? We think so. But there's a lot of execution behind taking that and getting the right things in the marketplace.

ED LAZOWSKA: So let's pursue this a little more. Steve has lifestyle and work style, and for the three of us and everybody in this room, those are unfortunately pretty much indistinguishable. That's what my family tells me, at least. So can you cite some specific benefits, or transitions, say from the Xbox side to the corporate side, or something like this? This is part of what you want to take advantage of, certainly.

BILL GATES: Well, certainly electronic mail over time is something that, on your mobile phone, you want it just to be there; you don't want to have to think about synchronizing. You take a picture with that camera and you want to, just based on your calendar, you’d like it to be categorized. Is that related to some work project, is it something that ought to go off to the relatives in a very automatic way? If there's a message of great importance, you might even want to be interrupted while you're watching TV, and so you're IPTV screen scrolls a nice little message down at the bottom saying, you know, a new urgent piece of mail came in from a work colleague. And so thinking of what's important, what kind of notifications do you want, how do you want to communicate. Take schedules, this is what we're working on. Ray and the teams are pulling things together to bring this to a new level. You'd like a work calendar that's just got your work items in it, and a family calendar that has some connection to that but it's not exactly the same. So making that work across your different contacts, and across all those different devices, if you do that, make it very simple to use, everybody will just say, wow, that's common sense. I like this Microsoft integrated solution.

RAY OZZIE: All right, I actually think that there's also a tremendous enterprise benefit to this—the fact that people generally use the same tools at home and at work—that really comes down to reducing the TCO of a desktop within the enterprise. I mean, many, many kids learn how to use Word, how to use Excel, things like that, at a very young age. And people use these Office productivity tools in many, many aspects of their life. This means that it's one less thing that an enterprise has to train their users in when they get to the work environment. The same is true for Outlook mail. People frequently use Outlook in a broad variety of situations, and when they come to work, and they're using that same Outlook user experience against the Exchange servers it really does help to effectively reduce TCO.

ED LAZOWSKA: So let me pursue your contribution to the company for a second. As I said, really since your time at the University of Illinois in the late ‘70s, you've been doing productivity and interaction and communication, and maybe people here aren't fully familiar with the technology behind Groove. I wonder if you could talk about that for a minute, and then, Bill, if you could talk about how this fits into Microsoft's plans.

RAY OZZIE: Sure. You know, I've found collaboration and computer-supported cooperative work to be something that I've been really passionate about for probably twenty years, and when Notes came out in the early nineties, it really was coming out in an environment that I'm sure many people out here remember. It was the re-engineering of the corporation era, and people were trying to use technology to streamline how people worked across boundaries within an organization, streamline business processes within organization boundaries.

By the mid '90s, we started noticing at Lotus IBM that the more leading-edge companies were doing something that's very commonplace today: They were trying to use technology to extend those business processes outside their company boundaries. And the more that I looked at it and that we looked at it, it really looked to be a basic structural change that was happening within business, and the people who were trying to do it using our technology were really struggling, because essentially the technology that we had built for them was really focused on enterprise IT deployment, datacenter deployment, centralized management at the enterprise level, and taking those same tools and trying to tell your partner to use those same tools when you're trying to get a network of companies working together in more of a mesh was really a nonstarter.

And there are two real ways that you can really view this moving forward. One is that it's a tremendous professional services opportunity to go out and help each company in a parallelized fashion try to get these things working together or to take a fresh look at the architecture of the product to see is there something fundamental that we can do using software to make it—in a lower-cost, higher-scale manner—possible to securely get these companies to interact with one another at the business unit level, not just at the corporate IT level.

Groove was really essentially taking a fresh look at the technology, thinking from more of a peer-to-peer standpoint where a business unit within a corporation could decide, oh, we've got a project, we've got some kind of a task that we need to do involving these eight companies, let's download it, let's get these other people to use it, and get the problems solved very, very quickly, without boiling the ocean across enterprise boundaries.

The way that it has turned out, it's been tremendously powerful in terms of balancing the use of centralized and decentralized systems; that is, what we've found in almost every Groove installation is that a company had embraced an internal information management technology such as SharePoint throughout their organization to manage content in ways that are very appropriate from a control perspective, a compliance perspective, at the team level or at the enterprise level, and then when they have dynamic processes that they need to involve people in the outside, or they have highly mobile workers who work on the outside of the firewall a lot, they would use Groove to take things out of the centralized repository, use them in a work process, and then have them either automatically or manually populated back into those centralized systems as a result.

So I think, I mean, what has excited me about coming to Microsoft is the fact that having worked in collaboration for a while, there are different modalities and different usage styles that reflect really the way that we work together, both as individuals and organizations, and that Microsoft has made significant investments, both at the real-time modalities around Live Communications Server and Live Meeting, in asynchronous collaboration around Groove, in very robust enterprise content management and portal with regard to SharePoint, and, of course, Office, which essentially ties into all of those.

ED LAZOWSKA: So technologically, peer-to-peer, it's a business unit scale.

RAY OZZIE: That's right.

ED LAZOWSKA: There's obviously a big security component.

RAY OZZIE: There's a huge security component. In order to build a system that you can trust across enterprise boundaries, you really need to build in security from day one; it has to have the ability, and Groove does, to essentially federate with the directories that each organization happens to be using, because you can't expect that someone else is going to necessarily, when you're collaborating, make the same technology infrastructure choices that you did, or even be on the same version of that, so it very dynamically adapts from a naming and directory perspective.

Everything that happens in Groove is automatically encrypted; we call it complacency-immune security because the reality is, to be secure it can't require users to turn knobs and dials to get it to be secure. And a lot of Groove's success was in government markets—particularly post-9/11—in enabling government agencies to work together more effectively, and government and NGO, particularly in the humanitarian realm.

ED LAZOWSKA: Got it.

So I want to get to how this fits into Microsoft in a second, but first I remember five or six years ago being out at Woods Hole with you in a meeting, and you were talking about the programming and development style in Groove. This is just when it was you and 20 people or something like that. And the two things I remember you said are, first of all, that software is always legacy software, that is from day two you're working with an existing code base and evolving it, and it's not that you define the spec and then you implement for a year and a half to the spec and then you look around and see where that is.

And the other thing you point out was that it was sort of event-based style of programming rather than traditional procedural style of programming, entry point, whirr, whirr, whirr, exit point.

And that was I think probably a pretty new culture six years ago, whenever we had that discussion. What do you see in Microsoft along those lines today in terms of the general approach to development with the team?

RAY OZZIE: Well, interestingly there are common roots in a lot of things that I see at Microsoft. A lot of the people that started Groove came from Iris where I did Notes, and some of those people came from DEC, which has a common ancestry in Dave Cutler, and a lot of those same people are at Microsoft. There are a lot of engineering disciplines that are very much in common.

ED LAZOWSKA: Cutler is kind of a spec guy, though.

RAY OZZIE: There are certain styles of programming disciplines around common things, structure, and build procedures, and things like that, that are required in order to build large, scalable, complex systems, and we've been very fortunate in having been exposed to some of those, particularly at an earlier age.

One of the things that I have been very proud of in terms of Groove—in particular in terms of getting a lot of very complex code woven together in a way that just works for the user and the customer—but one of the things that I've been very, very surprised and pleased about in coming into Microsoft is that Microsoft has made huge investments internally in security in terms of tools to help flesh out where the problems are. Attackers out on the Internet are beginning to move from the systems level up into the applications level, and I believe that in the process of building Windows Vista, a lot of tooling was built in order to make it an incredibly resilient system so that people could have confidence that what they were installing was not going to be hacked into.

A lot of those tools, the moment that we were acquired by Microsoft, were available to us. They're in a beta form, ultimately they're going to be in the next version of Visual Studio, but they do static code analysis so that they look at your code to say, oh, here's a buffer overrun, here's some kind of a bad error path that would ultimately result in a vulnerability. It's been a tremendously positive experience to have those made available, and I think that the partner community is going to benefit a lot from that.

ED LAZOWSKA: So this is great, so this is another thing I wanted to touch on. I saw this because I'm engaged with Microsoft Research, and a lot of those tools came from Amitabh's group a number of years ago, and have now moved into the product organization. It's a great example of sort of tech transfer into the company and ultimately to customers from MSR, but it's interesting validation that a state-of-the-art development team from outside the company gets access to these tools and, gee, look what they do for you.

RAY OZZIE: Yeah, it's real, too. I mean, the vulnerabilities—and whenever you have a team of, in our case it was 60 developers, roughly 60 developers doing 6 million lines of code over a small number of years, people are focusing on things that are customer-relevant, they're trying to solve a problem for customers, and even with best efforts there are going to be problems in the fact that these investments can result in automated tools that can help on a very mass scale—secure things is a big deal.

ED LAZOWSKA: So, Bill, can you talk about these tools a little bit, both what they are, and what they're doing for security in Vista and other products, and what the impact is going to be on customers? Again, I've seen these come down the pipe over the past five or six years, and it's really remarkable. I think many of the things that Amitabh and his group did were to make approaches viable that people believed weren't going to be fruitful; it's been from the point of view of the software engineering community really remarkable what they've accomplished.

BILL GATES: Yeah, certainly when I dropped out of Harvard, right at that time there was all this talk about proving programs correct, and actually understanding properties of programs. And for things that were, say, 40 or 50 lines long, particularly sort programs, they got really good…

ED LAZOWSKA: We had that nailed.

BILL GATES: …and now we have these systems that are often tens of millions of lines of code in size, and where you have somebody with evil intent who in a fairly invisible way, because the Internet brings all these computers onto a single network, is actually probing for any mistake that might lead to an escalation of privilege.

And in order to deal with that, you've got to have programming the type systems at a higher level so that you're less prone to mistakes, and you have to have this deep analysis.

And so those tools literally over the last three years have gone from being purely research tools to something where today, when you check your code in, before it gets into the check-in tree, you have to run this thing called PREfast over that code, and so you don't even get it into the tree if it's not following these conventions.

When we're designing new interfaces, there's literally this thing called FxCop that goes and looks to see are you doing the types the right way, are you passing the links the right way, so you can't make a whole class mistake there.

The whole security front has required innovation in many respects. How do you isolate these systems? A thing called IPSec and deep advances that go beyond this firewall approach that people thought was good enough, things having to do with quarantine and inspection and applying group policy, using virtual machines; that's going to give us deep isolation.

Code vulnerabilities, we need these tools that we've talked about, and also this ability to update things and know which updates should flow right away, which ones should go just to the Internet-facing systems, which ones should go to all systems. And so getting that infrastructure in place, what we call SMS, the whole Microsoft update capability, that's been an absolutely huge thing.

Another leg of this is authentication. As Ray said, people seek the weakest link in the chain, and so you have a whole stack of software where they can be probing, but you also have this notion of who's on that system. And so if you're using passwords, that can often be the weakest link. If you just call an IT department and say, “I'm Joe on the road, reset my password," and you're not Joe, usually that works. Or if you create a Web site that people have to enter passwords into, you'll find they enter the same password that they use at work, and so immediately you have access to their…

ED LAZOWSKA: I worked for a company once 20 years ago that required that you change your password every month, so everybody's password was APR84, MAY84, JUN84. It was no problem figuring those out.

BILL GATES: Yeah, passwords are never going to be as a single proof point good enough, and so getting in so you can get particularly smart cards and various biometric things is important there as well.

And there's a big investment that's been made in these directories, and that's actually a big strength of Microsoft is the Active Directory capabilities. Now, making that higher-level so it connects up to all the applications, works across these multienterprise scenarios, so-called federation capabilities, those are things that have become top of mind because of the security focus in our customer base.

And if you look at our whole R&D effort, security would be the biggest thing of all the different activities. And I'd say since that happened, which is about three years ago now, I'm really pleased with the progress we've made. We were in the dialogue with customers about exactly how they know that they're secure, and what the key infrastructures are to achieve that.

ED LAZOWSKA: Let's go back to Groove to talk a bit about how it fits into Microsoft's plans and what it brings and what Microsoft already had on the table that Groove fits into.

BILL GATES: Well, you can go back several decades and say that Steve and I would often sit down and say, hey, of all the people, if we could hire one person who's not at Microsoft, is there any way we could get Ray to come. And so whenever he would go and do something new, we'd always see, and finally this time we found a great combination where we could take not only the work of Groove but also of a pretty incredible team there and add that to our Office group.

The idea that people are working in new ways, and the business benefit of making people more productive when they can get at information in a simple fashion, that's actually one of the biggest opportunities we have.

When you look at Microsoft's portfolio that Steve talked about, some of them are areas where we have lots of competition: search, video games. Some things that are actually the biggest opportunity for us I think people miss because actually there aren't companies investing in the R&D and coming in every morning and saying, okay, how can scheduling be better, how can meetings be better, how can information flow across these boundaries, business intelligence, how can that be better.

And so we've said that through Office on the client and these Office products, particularly SharePoint on the server, we can revolutionize how information flow works inside companies, but we can't do that by saying it's just the one company and only their trust structure.

And so, as we looked at that—the idea that we could bring in this Groove peer-to-peer capability—it's always been about the thing we believe in most, which is letting people, bottoms up, be empowered, and the minimal amount of bothering the IT department as possible, from a cost point of view, from a complexity point of view. If they can just put in place a few things to let companies have federation, then these employees with no touch, no manual touch at all, are creating this ability to share. And so, as Groove has been integrated into our Office effort, it's not just that peer-to-peer capability. We're going to show people how that fits in, in this magical way, with SharePoint, and show people the scenarios where these things come together. Ray, in particular, is going to take the entire productivity thrust that we have and make sure that we're driving that forward. He’s always had great ideas about collaboration, and software today still falls way short of what's possible there, and we'll embody that in new generations of Office. Of course, Groove will keep evolving. We have got a great development center out in Massachusetts where that work will go on. But it will fit into that overall Office strategy.

RAY OZZIE: I should add one thing also. A lot of people have talked about Groove as being peer-to-peer, because that is the fundamental unique aspect of its architecture. But one of the things that really isn't widely talked about is the fact that when you're trying to accomplish these meshed scenarios of a group in one company working from behind one firewall to another, or a customer using it from China to coordinate their development with San Francisco and India, the fact is that it requires a mediation of a service, and there is a service underlying the Groove infrastructure. I think one of the things that portends is the fact that a lot of the productivity software that we see and that we've used for many, many years can essentially come alive in some way, shape, or form by bringing the benefits of having a services infrastructure out on the Internet together with these rich user interface capabilities that people have grown to expect—and should—with the power of those devices, and the combination of that software and those services as taken from product to product, both in the productivity side of the world and in things like Xbox. Really, it's a tremendously exciting future in trying to understand how those services evolve.

ED LAZOWSKA: Just to state the obvious, it's pretty clear that this workgroup level is a sales and marketing advantage, too. You don't have to convince the corporate IT person, the corporation as a whole, that they should embrace the tool.

RAY OZZIE: Many of these things start out grassroots. And in order for collaboration, in particular, to be successful in communication programs, the ones that are the most successful are the things that people really personally can relate to and understand. But, ultimately, they all have to be IT-relevant because we live in an environment where compliance is a very important thing. And so, ultimately, you really have to find the right combination of the center and the edge of IT management of things that people really know how to use intuitively, and really like to use at the edge.

ED LAZOWSKA: I have a very naive question. Is Microsoft in the business of helping companies with security compliance? There's been a trend recently towards appointing chief information security officers, and having an IT-capable person on a board.

BILL GATES: Absolutely. The software solution we built around Sarbanes-Oxley really let us show off what are core productivity features: information sharing, keeping people up to date, how you do notification. But taking it in the very urgent projects people have, then you see the exposure: “Wow, SharePoint can really do that, Rights Management can really do that.” People often say to us, how do we deal with all the e-mails, because we're very a strong e-mail culture company. We talk about how that's been balanced out with more—you have e-mail, a lot of it will be notification about a SharePoint or a workspace, where people are collaborating together. And then we show how we take that and allow you to have policies about retention. Do you want people to keep certain mails, do you want people to clean up certain e-mail? The software can enforce those policies.

This Rights Management idea that when you send a piece of e-mail, you can say only a certain scope of people should have that forwarded and have access to it—that's, I'd say, a very popular capability in electronic mail because if you're using it, say, for financial data before the close of a quarter, it lets you have some traceability that you've done what you're supposed to do to control the information. And then, what we do with the document life cycle, it means that if you're supposed to go back and find information of a certain type, you can really say, we've absolutely found all the documents that are there, and enforce policies so there are only the documents you want to have the storage for.

ED LAZOWSKA: We've already covered a lot of this, but it's pretty clear that Groove and Ray bring a set of things to the table. There's a systems aspect, sort of peer-to-peer, there’s a security aspect, there’s a collaboration and user interface aspect, there's a services aspect you've described. They're going to spend the afternoon hearing about Office and Vista. Can the two of you talk a little bit about what we can expect, maybe not just in the next year, but extrapolating a little bit in terms of fitting the applications and systems together to support this kind of collaboration?

BILL GATES: Yes. Groove actually had to build a lot of sort of systems infrastructure and applications infrastructure. And one of the things we can do here is get a lot of that systems-type capability built into the Windows platform. For example, a lot of the time this summer’s been about how do we make sure that the peer-to-peer capabilities in Windows Vista are going to be in exactly the right platform for what people want. How do we make sure that the way we're looking at these security things gets down more into the platform? How you keep track of the people you trust, so that the chain of trust is there, how do we do those things?

And so, we always try and make sure we're drawing that boundary. Sometimes the applications will build something that gets popular, and then it's clear—like OLE as an information-sharing capability, we moved that back down into the platform. So we've got to keep the pace of taking those collaboration capabilities and getting down into the platform, so not just Office, but all the software vendors who build on the Windows platform can do unique things there.

RAY OZZIE: The thing that probably excites me the most is when I look at the development of what's going on surrounding Office. Sometimes when people think of Office they think of the software of Word and Excel, and that, but Office has really evolved into a much larger system of software, of servers, services, and solutions—that as an umbrella we'll refer to as the Microsoft Office System—but in essence it takes a much broader view and says that the things that you do as an individual with Office relate to communications, number one. You generally don't build spreadsheets and documents just for the printer anymore. You deal with other people, in terms of co-editing cycles, and you’re working on an RFP or a proposal with someone else. The combination of Microsoft Office, of SharePoint, of it with Groove, of it with Live Communications Server, and Live Meeting, really brings it up from the perspective of business value, from personal value to business value. And in terms of bringing the value from the perspective of just clients software to what you can do when you bring that thing together with servers within the enterprise, and services outside the enterprise.

So some of the things related to document life cycle, and actually some of the more complicated things related to compliance—right now those things are being done primarily by large enterprises who have made significant investments for a good reason, in making sure that they've got their processes and tools in line to do that. Those represent opportunities to bring those same kinds of things in terms of services out to a much broader set of organizations who might find value in those document life-cycle types of functions, but don't necessarily have the investment or expertise in building those systems.

ED LAZOWSKA: Let's go back to Steve's talk and tie this to it. Steve said Microsoft was going to grow in three ways. One is expanding the anchor businesses, another was sort of developing new businesses—that was the blah, blah, blah part of his talk—and the third was delivering software and services. When he talked about software services, it wasn't becoming a services company, or at least exclusively, and it wasn't software as a service, it's sort of software and services together. Can you talk a little bit about the future there?

RAY OZZIE: Again, coming back, when people say software as a service, in many cases they immediately think thin client, no software, Salesforce.com, you know, type of model. For many types of applications I'm sure a browser-based approach might work—we all use browsers against a variety of Web sites—increasingly, the capabilities of Web sites using DHTML technologies are getting more and more powerful.

However, we really believe that there is—I, frankly, believe in the power of software, on mobile devices, on PCs, the capability of these devices in terms of expressing a rich user interface, a human interface; each device that we carry around—the laptops that many people here have huge disks on them. You have lots of pictures and media and documents, and things like that. And essentially those systems can work very, very, very synergistically with services that are in the cloud.

So when we talk about services, essentially we're saying, how can we bring life to that software by rethinking them in this era where there is a tremendous amount of public information, where there are services up in the cloud that can offer enablement of communications, of storage, of media services, transaction services, location services, things, as I was saying before, in the document realm there are lots of—when you bring software together with a cloud-based services infrastructure you can end up with something that's much, much more powerful than either one of those individually could be.

BILL GATES: Yes, the service concept, we'll see it at the consumer end with things like Xbox Live, where signing up for contests, and being a spectator, staying in touch with your friends, that's something you're going to want to be part of. So video gaming will have gone from being just a thing that you do in one location, with just the people there, to a far more social thing, that opens up new game genres. So it revolutionizes that business.

If you go into the realm of business activity, there are so many things that have been complex in terms of setting up servers, that many times you just don't want to do it. We can say, hey, the virtual equivalent of that server is available for some type of yearly fee. What it means if we have to think of our architectures as being deliverable in two different forms, as a piece of software that runs inside the corporation on the server hardware they buy, and as a service that we go out and offer to them. And we've been coming up the learning curve on this. We've done quite a bit with e-mail. Steve mentioned we've done some acquisitions around people who are working in this way.

One form of this is that we'll have storage in the cloud. So that you can connect up with different devices, instead of actually having to have those devices be together to synchronize, you'll simply connect up to the cloud storage we've created and the information will flow. A lot of that can be very automatic, having your favorites go to your different devices, or the way that you use e-mail. We're just going to make that straightforward for people. And the technology trends are making it so we can offer those things at lower and lower cost.

ED LAZOWSKA: So a different subject. I've heard each of you speak about the problem with information overload, something that everybody faces, and I guess the obvious thing is, our ability to absorb information is something that doesn't change, whereas the information we're subjected to increases relentlessly. And there’s the question maybe about whether this is too much information or too little information. I think of it as we're all sort of drowning in data and we have a very hard time extracting the knowledge from it. Can you talk a little bit about whether this software-plus-services approach has a bearing on that, and more generally what Microsoft is doing to sort of provide me with cognitive assistance to manage the data that I'm drowning in?

BILL GATES: The user today has to manually think about each of these devices. Even when you have multiple PCs, you're kind of in charge of saying, okay, what's on here, what version of the software, what patches, which files did I get on here, when did I synch this thing last, a lot of special commands. And when you go across device types it's even more complex, because the way that schedules are handled, or contacts are handled, will be very different on those things. In the future, that should just happen for you. The data will be there.

Another very important thing, though, is software getting a sense of the different things that change, which ones are important to you; that can be somebody trying to call you, and deciding whether you need to be interrupted in a meeting; somebody sending you a piece of e-mail, deciding where that should show up and how much you should be bothered to pay attention to that; or simply something happening on a Web site, a stock price changing or a flight arrival time changing. As you're navigating that information, your ability to subscribe should also include an understanding of how important it is and how you like to be notified.

RSS as an industry standard has really started us down this path of saying, okay, when information changes through a variety of pulling and pushing techniques, I should be able to have that come onto my machine. So if there's a new site that I like, when it has new articles, I'll see some indication of that and be able to go to that, so it's not just e-mail or trying to go look and see Web sites and see what's changed.

The infrastructure around that today is way too limited. We don't really have a notion of what your relationship is with various people, and therefore how important their messages are. We don't really have your digital schedule in such a form that understanding your context is quite automatic.

Over time—and Microsoft Research has been looking at this—as we add things like a microphone and a camera, even knowing, are you meeting with somebody in your office, are you paying attention to your computer screen, working on something like that, a lot of those things ought to be implicitly available so software can behave in an intelligent fashion.

RAY OZZIE: One of the other interesting things that I've found and I've had the pleasure of working on in the short time that I've been here is that Microsoft is very uniquely positioned because many of the tools that you use in your life, or could use, can be woven together, because we can accomplish scenarios across these different devices that we might not be able to do if they were not all done by one vendor.

So, for example, many of the things that I do on a daily basis weave together my use of Exchange; my use of Outlook as the user interface for those messages coming in from Exchange; my use of MSN, because my wife and family use MSN, and I actually would like to aggregate some of those things into my Outlook user interface and share some things with her; and my use of my Smartphone, where, you know, today, when I'm here, I didn't bring my laptop, but I'm able to actually look at my messages and my calendar and things like that.

It's all unified through a server and services infrastructure. And because we have the ability to map out the scenarios end to end amongst all those different places in my life—you know, work life and home life—we're able to accomplish things for users that we might not otherwise be able to do.

ED LAZOWSKA: So these are things only Microsoft can do in some sense.

RAY OZZIE: I believe so. I believe so.

BILL GATES: There are really two levels. There will be a level of interoperability you get through industry standards. And we're going to keep driving those standards forward, like RSS, like Web services. But when it comes to the user interface, exactly how repeating meetings get done and the photos and the contacts—some of these messy details we'll be able to task that user experience across all the different devices in a very deep way. And that's part of the job Ray has and I have, to make sure that we're thinking of that in terms of the user, even though it's multiple Microsoft groups who are building the technology.

RAY OZZIE: I mean, at this point in the industry there's no shortage of technology. People have many, many, many choices. And they have been gravitating clearly in the market, particularly on the consumer side, to things that solve the whole problem. They're thinking about solutions. I mean, this is true in IT. This is true with consumers. If it just works, if it's woven together in a way that's easy to use, that's easy to manage, that's the product that wins as opposed to some other form of technology.

And that's why it's scenario-based. Mapping out those administration scenarios, those user scenarios, and weaving together hardware, software, services in a way that makes it all just work very seamlessly, I believe, does give Microsoft a competitive advantage and will make those devices and pieces of software much more useful and valuable and less costly to operate than when someone has to weave them together themselves.

ED LAZOWSKA: So let's talk about the role of search and the future of search in this context of information overload and information management. You know, Bill, you've said that we've just barely scratched the surface in terms of what search can do, and I'm sure that's right. But on the other hand, it's changed the way I operate completely. I don't have a Favorites list anymore. I just search for the same thing day after day after day. I mean, I type in "New York Times" and there it is. You know, I don't pull down a Favorites menu. I've long since given up trying to organize e-mail messages and organize files.

You know, when I think about the desktop metaphor, I look at my own desktop. You know, it's just an incredible heap of paper, okay. And, sure enough, in the fullness of time, my digital desktop got to look just like that as well. And search has sort of saved me.

So what do you envision in the future? Obviously personalization; maybe you could say a bit about that. But what else? Because this seems like such a powerful change in terms of how we operate and reducing the amount of time we have to spend attempting to organize things so we can find them.

BILL GATES: Well, computers are very good at taking all the things you've done and being able to build these quick search indexes from those things. One thing that Microsoft will be educating people on is that search is a very broad thing. It's the documents that you own. It's the documents inside your business. And then, of course, there's the entire Web.

And in each of those pieces, you want a very simple interface. If I'm inside Word navigating a document, I want a nice search interface. If I'm looking in the directory trying to find somebody with some expertise, I want a nice simple interface.

Inside the corporation, it's fairly complex in that there are rights boundaries. Some people should be able to get some information and not other information. And, in fact, the idea of what document is most important, the way you figure that out in the corporation is completely different than you do that on the Web. Some people have tried to apply those Web technologies in that business realm. And the results in terms of the quality of what they get is actually very low.

Part of what we're seeing in innovation now is taking the world of structured data, like a directory or a customer relationship management system, and the world of documents and bringing those together. So that in a business, when I say, "Who manages this customer?" that may go get the answer from the CRM system, not trying to go through all those documents. And so we don't want to divide the structured world and the unstructured world. We want to bring that together on behalf of the user.

In the next generation of Office, what we call Office 12, we've done that. We've built the search capability in so that it gathers up all the information. Even from your accounting software, you can navigate in and find the records that make sense.

And so we're innovating on the desktop. We've got an add-on for that today that's free. We've got it built into the Windows Vista user interface in a very deep way. We're innovating at the corporate level, and that's an extension of SharePoint in the Office product family.

And I really think it's going to get people to think about how they organize information as a company. We've got individuals thinking about how they create documents. But the way information flows inside a company, there are too many tools and interfaces today. It's done mostly with e-mail attachments, and people don't really have document management because it's too complex and too expensive to get for everyone, so they got it just for a few people, but then those people need to work with everyone else, and it didn't come together.

So the same revolution that took place on the client side with Office becoming common sense, we think in this next round that'll take place on the server so that businesses think, "Wow, the way that I manage data, the way that I protect it, work-flow it, share it, is this Office SharePoint capability.”

Now, the Web search is also very important. That's where the advertising piece comes in. Today that's a treasure hunt of links, and it doesn't really know when to implicitly include your current location. That's sort of a qualifier. Clearly if you type in things like "pizza" or "police" or "doctor," maybe it shouldn't give you the one in Florida; whereas when you type other query terms, you're interested on a global basis. And if you know about a subject, the type of Web sites that should be brought to you are very different than if you're exploring them for the first time.

So watching your history, knowing what types of document things you're working on, actually not just indexing the documents on the Web but understanding those documents, that's going to turn search into something where you actually will get answers in the future. And so we feel very lucky that even though there's an aspect of this, the Web piece, that we're playing catch-up on, what would exist today is really quite poor compared to what we, or perhaps others but certainly we, can do in the next two or three years there and really redefine search, so you'll look back on today's search and say, "Hey, that wasn't that good."

ED LAZOWSKA: Got it. Let me talk about R&D for a bit and Microsoft Research. I think many people don't understand that while every engineering company does R&D, at most engineering companies that's engineering the next release of the product or looking maybe two releases out. Microsoft is one of a very small number of modern information companies that makes a significant investment, largely through Microsoft Research, in looking more than a couple of product cycles out.

So Microsoft Research is now—I don't know what—700 heads or something like that. And they are, from my point of view as a university computer science researcher, the best computing research organization in the world. They're doing what universities do, which is looking five and ten years out and trying to bring about fundamental changes. To me it's, in some sense, remarkable and rewarding that Microsoft has made this investment, because it puts you in the research community driving the entire field forward and creating what's essentially a public good, which is ideas that are in the literature.

How does this pay off for the company, and how does it pay off for customers? You've given some examples; for example, the work that's led to improved security. But there are surely lots of others.

BILL GATES: Yeah, I can say with total confidence that the best investment the company ever made was building that research group. It's been a phenomenal experience for us. And I think other companies should be doing more of the same. In fact, in our field, because entities like Bell Labs and Xerox and even the Japanese, who had made a thrust at one point, have really backed out of the scale of research they used to do. If you take it in aggregate, there's less research today than ever. And yet the really interesting problems—vision, speech, ink, security, learning—all these things, we're going to have big breakthroughs on in the next 10 years.

ED LAZOWSKA: We'd better.

BILL GATES: Absolutely. And so when we started Microsoft Research, there were two things we were afraid of. One was that our research group would be very parochial and that, say, they'd have an approach to document translation or speech recognition, and that we'd lose the ability to know what was going on in the universities, because universities are key to this.

You know, if you take computer science research as a whole, most of it goes on in universities. There are aspects of it that have to go on in companies, including sometimes taking what's done on a smaller scale and combining it with other things so it can be productized, and sometimes doing unique things. And so we wanted to make sure that it enhanced our relationship with the universities.

So that was a boundary we had to work on, and that's gone incredibly well. In fact, it was just a few weeks ago we had top faculty from all over the world here talking about what they thought our research agenda should be, talking about curriculum, talking about policy, and all the different issues that concern them, and really viewing ourselves as part of that community in a very deep way.

I had no doubt that Microsoft Research would be able to hire brilliant people, and that's worked out better than I would have guessed in the number of papers that are accepted, the breakthroughs they've made. But the second boundary that was key is that between our research group and the product group – you know, Xerox PARC was famous for not having a very good relationship with the product groups. The one product that came out was kind of a flawed product, the Star, and didn't go anywhere. Here –

ED LAZOWSKA: Sorry. Let's remember that the laser Xerox machine came out of Xerox PARC.

BILL GATES: Well, okay, yes. I always think of – you're right; it was a part of PARC that more than paid for the whole thing many times over.

ED LAZOWSKA: Gary Starkweather, who just retired from Microsoft Research.

BILL GATES: Incredible group of people. So the computer science part didn't take it quite the way that it should have. Well, here we've managed to not have any problem with that at all, because the people we hire want their things to fit into products, and we've created a number of things like a fair where people show off their latest breakthrough ideas. And it's just created this very positive relationship between the researchers, even though they're in many locations. You know, I have a clear sense of what's going on in Beijing, and I see product groups taking advantage of that.

So whether it's the new version of Windows or Office or Xbox, the way Xbox does grass and hair, it wouldn't be the same without these breakthrough things, just hundreds and hundreds of ideas. And we can turn to Research and say, "Okay, we're trying to build this big file system and database in the sky." And Ray and I are in a meeting tomorrow to go through exactly how the researchers are advising that we do that. And it's going to be an important asset.

So other than probably IBM, we're the only ones doing large-scale research. And I think we get—and I encourage others to do it—but I think we get a huge competitive advantage by taking that long-term view in a way that others are not.

RAY OZZIE: And, actually, the very first question you asked about—you know, as the newcomer what did I not expect to see—I have been amazed at the number of projects. I know of four very, very specific ones that I've been exposed to where the researchers are working hand in hand because the product groups have decided that they needed that technology that these people really were pioneering in, and they're working, the transfer is working extremely well.

ED LAZOWSKA: Yeah, I've been to a number of the tech fests, which are this sort of marketing of the product groups, as you talked about, and it's great. There are hundreds of booths, and each one has a little place where you walk by with your Microsoft badge and you’re suddenly on the dear friends distribution list for that group, but it's a very proactive form of technology sharing and really works very effectively.

Do you want to say a bit more about university research and the federal government's role? I think a natural question is, gosh, Microsoft is putting hundreds of millions of dollars a year into research that's looking forward five and ten years. What's left to do? What's the role of universities, what's the role of the government historically in going forward? How does that pay off for the company as well?

BILL GATES: Well, I think the leadership of the United States in information technology and in biology really comes from the huge R&D investments that the U.S. government puts into the university system. The basic advances, most of those, are going to come out of academia. And the boundary in those two industries, between the university and the corporations, the U.S. has handled that better than other countries, and so the give-and-take, the push towards relevance, has been fantastic.

There are some scary trends here where the amount of federal funding—and there are many ways to measure it, but almost any way you look at it, the character of the funding, the amount of the funding, has been actually on somewhat of a downward trend. At the same time, the equivalent funding outside the United States is going up somewhat. And so it does raise a pretty interesting question. All our university colleagues are saying that this is presenting the challenge for them in terms of the size of their departments and the kind of work that they can do. And so it is a bit scary. And it's almost ironic—given the entire Internet and so many of these technologies—that there ever was a vivid proof that allowing people to work on things which are fairly unstructured, and you don't see a clear path between that work and the actual products that it can pay off, that we've had it here in the United States. And so the jealousy of what went on here is partly what's causing other countries to try and step up.

Now, I'm not saying it's going to be easy for them to do that, but then again it certainly is something that we ought to make sure we're pushing to a new level as they enter the game.

ED LAZOWSKA: I think one thing to point out is that Microsoft research gives you a link to that work, and in some sense a competitive advantage in absorbing it into the company as well.

BILL GATES: That's right. The relationship with these universities, not only did we have to make sure our people were open to it—and they'd say, hey, I was pursuing approach A, but actually at this university approach B is better, perhaps even synergistic with what I've been doing, and we can go in and get access to that.

The universities are really great at wanting to talk about how they've set their agenda, wanting to see their things get into products. I mean, that's why this deal is moving as fast as it is. And I'd say it should be moving at a great pace, because we can solve the tough problems. All of these people engaging in digital work style, digital lifestyle, means that the impact—whether it be security, user interface, basic questions in artificial intelligence—we'll be able to apply those to the benefit of hundreds of millions of people within years of the advances being made.

ED LAZOWSKA: Well, we are out of time. I hope you found this enjoyable. Thanks for inviting me, and thanks for participating. (Applause.)

END

Due to the varying sound quality and subject matter of tapes, the information in this transcript may contain inaccuracies.