Bill Gates, Microsoft Corp.s chairman and chief software architect, sat down with eWEEK Senior Writer Darryl K. Taft to share some of his vision on where the industry is heading, whats next in Microsofts sights, and the Next Big Thing for software developers. Gates granted eWEEK an exclusive interview at the VSLive conference in San Francisco last week, where he keynoted on his vision of Seamless Computing. Gates spoke with eWEEK on the day the against the company, but Gates deferred questions on that to his attorneys.
On your most recent “Think Week,” what were some of the themes or things that you came away with?
Those are in part probably the funnest part of my job, seeing some of the research work coming out of universities, getting very up-to-date on the research inside Microsoft, and then Ill pick some areas of things I need to understand, like where are microprocessors going.
This last [Think Week] was my time to really study some of the wireless technologies, ultra wideband and the so-called WiMax 802.16, to try to get a sense of who the key players are, what the new capabilities will be and how we should factor that into our software design. Theres a thing called mesh networking, which is software making all this stuff work together in a way that lets you do video and audio in a pretty neat way.
And this idea that were going to unify on the network voice, video and data, Id say that came through as a real theme of my Think Week, thats some of the work were doing, and some of the phone companies are now moving toward that unified network. Thats another one of those Holy Grails that you can probably find in an article back when eWEEK first came out talking about unified networks. But thats really coming to the fore.
I spent a lot of time on security simply because making sure all the groups at Microsoft have a real integrated strategy and clear road map around security—that was very key to me. And getting people to understand the nature of the problem, that isolation is a very, very key tactic and that we really have some wonderful isolation tools. Today, people think that is just the firewall around the perimeter, but theres a lot of things around quarantine and individual PC firewalls and a thing called IPSEC [IP Security Protocol] that go into this isolation that are going to give people a lot more control.
So, even when a contractor comes on-site around the corporate network, that they dont necessarily have the privileges. Just thinking of [security] as that one wall, unfortunately, that is inadequate. So, I ended writing up some things about that.
There are always some surprises for me. Our progress on things like automatic machine translation is one. I [read] four or five papers on that. And I knew Im funding those guys and I like their stuff, but theyve really made some progress and they showed results. Theyve taken a bunch of our technical-support articles and they had half of them hand-translated and half of them machine-translated into Spanish. And then they would have users rate their satisfaction with the articles to see if there was any correlation between the machine ones versus the human ones—and there was no difference there.
So were looking not only at using that ourselves to translate more things and save costs, but also, what partners do we reach out to to productize some of those advances? And its always neat to have something like that where you go wow, that team has really stuck to what they said they were going to do and made some real progress.
So the last [Think Week] was just last month?
Yeah, some time in February. I go off for a week with no interruptions. In fact, this one I was pretty religious about not doing e-mail, no phone calls and just day and night, other than sleeping, Im reading. One thing that was really amazing to me was I used to read the articles either on paper or on screen and then write a commentary. Now, because I have this big LCD screen and Ive got ClearType, I had the article in reading mode, and as I was reading I was typing the comments.
And the average number of comments I wrote per paper were over double when I was commenting while I was reading, versus reading it completely and then commenting. And the fact it was up on that one screen—a big screen that today is still a fairly expensive LCD, but wont be in the future—that was surprising to me how much of a difference it made to have that nice user interface.
Next Big Thing
Id like to get your thoughts on what the next big thing for programmers will be.
I think the big thing happening right now is the revolution that XML and Web services have generated, where even if software is running on a different machine somewhere else on the globe and even running in a completely different environment, we now have this structured way of describing information and exchanging information.
So, we can think of almost any piece of software out there running on another machine as almost like a subroutine, or we can go out and call it and get its capabilities. And thats a real mindset change, to think, OK, the information is structured and we model the schema, were very explicit about the format of the information. And any code thats anywhere that can manipulate this information, I can call it to do work for me. That is the biggest single thing.
And its a necessary advance. Some of the things people thought would happen when we got bit connectivity, Internet connectivity, actually needed this higher level of how software could talk to software before it could be real, like e-commerce. If all you have is presentation, HTML pages, you cant really do complex transactions. You cant track the state. You can do a consumer thing where somebody says I want that, I want to order it. But until software on the buyers side and software on the sellers side can talk to each other, then the dream of e-commerce is only partially realized.
And so, you know, Web services sort of had to be invented for these key applications. And the beauty of the architecture in terms of letting you talk not just to big servers but also to PCs and also to phones or whatever else comes along, because the runtime can be very simple, its very exciting.
Are you going to continue to lead the way in the standardization of Web services with IBM, or are you going to kind of take it the rest of the way on your own?
The only way that Web services applications will get to critical mass is if the customers can see that IBM and Microsoft and others have really come up with a set of protocols that not only look good on paper but actually have been tested between the products, and theyre actually at quite a rich level of capability, like the lower-level Internet protocols themselves.
And so its super important strategically to us, and I think to IBM and to others as well, to complete the rich Web-service protocols. The ones that are to get concluded this year include the ReliableMessaging, the security and the transactions. In fact, theyre very close to being done. Weve had a lot of what we call inter-op fests.
Like the one you and [IBM senior vice president] Steve Mills did in New York last year?
Well, he and I did a big event. That was a very big thing late last year. But even since then, without Steve or I, thereve been these inter-op fests. Other vendors besides IBM and Microsoft come in with their code to test against our code, and were really very concrete about, OK, is the specification clear about this piece, now lets get that to be very clear.
The inter-op fests have gone very well. And so its a pretty exciting thing that theres a matching set of protocols—for the low-level connectivity its the Internet protocols, for high-level connectivity were going to have better protocols than even homogeneous systems have had between each other. I mean things like reliable messaging in the past, you had to buy all sorts of extensive middleware and learn it and configure it. Here, native, essentially in every copy of Windows, were going to have those capabilities.
Web Services Standards
I look at what some of the other companies are doing, like Oracle for instance, and Sun. Theyre often on the opposite side of Microsoft when it comes to Web services standards.
Yeah, you often get a phenomenon in this industry where some companies are the first movers and that it takes overwhelming customer pressure to get the people who werent the first movers on it to join in. With Web services, we did reach out pretty broadly, and theres been a little bit of reluctance on the part of Sun and Oracle.
But even that, I think, you know, these are royalty-free protocol specifications. Now theres a fair bit of R&D work to do to build a platform thats got all these things and does it super well and has development tools around that. But this thing is a critical mass. And Oracle and Sun and everyone else, theyll be drawn into this. Theyre not the pioneers of it but, you know, once you get all the systems integrators saying, Wow, this is what were going to train our people on, and you get the pioneering customers, including people on Wall Street, doing these things, then its just a matter of time. Those pieces have come into place.
While were talking about Web services, last year I wrote a story that likened Microsoft to the Yankees, in that you were hiring up all the best XML and Web services talent at the time. Id like to know, whats your strategy around hiring and recruiting? And what areas are you looking to focus on filling this year?
Well, I think we are a little bit like the Yankees in that weve got a good track record, and sometimes people get a kick out of saying if we dont win, Hey, thats fascinating, why didnt they win? So, this is a great hiring environment right now in terms of being able to draw in smart, young people and show them the breadth of what we do and how weve got the staying power to really make the breakthroughs. Whether its making Web services real or speech recognition real or tablet computing real—very tough things. Each one of those things weve been at a long time, and were going to have to invest a lot more before they get completed.
A lot of our hiring is right out of the college level, so I did a tour recently where I went to U of Illinois, Cornell, CMU [Carnegie Mellon University], Harvard and MIT and really talked to the students about how these are really the golden years of computer science. That the hardware people are giving us these unbelievable platforms and that its really the software thats going to unlock the power of those things.
There has been a drop in kids going into computer science. So, even though Microsoft gets a good part of that pool of good people, were quite worried about kids in the U.S., less of them going into computer science, and less foreign students coming to the U.S. and joining these departments. And it is somewhat in contrast with the increase in computer-science enrollment in China and India. Right now, our hiring is very, very strong, in fact, and we have to make sure were picking the very best people.
And you seem to be. Thats the angle I took in the story, that everybody wants to play for the Yankees.
Microsoft Research I think has been the strongest group at Microsoft, in terms of benefiting from that. As you get very good people, other good people like to work with them, so Butler Lampson, Gordon Bell and Gary Starkweather: There are some people there who are magnets, and they create an environment where neat things are happening. And you know, we have groups that from time to time arent as exciting. We do lots of morale surveys and polling to make sure. Like, are the groups being managed exactly the way they should be? But thats a skill set.
Were a software company through and through, so over the years weve developed a pretty good methodology for knowing. I spend a lot of time with the summer hires. They all come over to my house and we have a party there. Actually, now theres enough that I have four parties there. But all the people get to come once. You know there are a lot of data points we can get to make sure that smart people are hiring other good people.
Any specific areas that youre pushing more than others [for hiring]?
Theres a broad set of opportunities, so we usually let the person say what theyre interested in. A lot of people like to work on Xbox, phones are pretty hot stuff, going up against Google, and doing something thats broader and deeper than theyre doing. You know, people kind of get excited about that challenge.
Surprise as Strategy
So, is that mentality still alive at Microsoft? Not so much the ruthless win-at-all-costs mentality you sometimes hear about, but a desire to win in every segment?
Well, win in the sense that we have customers who really say, Wow, this is great. The win that is probably top-of-mind at the company right now is that the customer perception of industry security and Microsoft security is, Hey, these guys havent done enough.
And so the mindset right now is lets win by surprising people, on isolating systems to make them secure and updating systems to make sure they dont have any vulnerabilities. We really need to surprise people, and we have some great work going on that I think we will surprise people in that area.
Theres a lot of enthusiasm for that. When you see the first few customers who really see those pieces, and we make them practical to roll out. Weve always been a culture where seeing the software in use is part of the reward cycle. Part of thats competitive. You know if Apple does something thats particularly good, there are certainly people at Microsoft who say, Ahh! We can do better than that. Like they did iPod, and now were doing some things like this Portable Media Center that bring video into it.
You said youre doing some things that will surprise the industry regarding security. What did you mean by that?
Well, when I say surprise, what I mean is, our software-updating product is called SMS, Software Management Server; the latest is the 2003 version. When I say surprise people, there are no secrets here, but just having customers talk to other customers and say, Hey, SMS really solved this problem, that within a small number of hours we can deploy and update and understand that its out there, and we can even set a policy that systems dont get onto the network unless they have this update. And getting rid of security [issues], its like a dog that doesnt bark, it looms over everything.
So when I say surprise people, I mean that the pieces will really come together and that we can show people how to audit whether theyve isolated their systems the right way. They can audit whether theyre doing updating the right way. And its required a lot if invention on our part. These are not things that mainframes or minicomputers from the past had solved. And yet they have to be solved. XP [Service Pack 2] is part of that, the software-updating servers, and all the things were doing to show people how to get isolation in their environment and check that.
Whats a bigger challenge to Microsoft right now and going forward: the legal issues youre facing right now or security?
Well, security by a huge, huge amount. When I have meetings with customers now, there might be one or two topics at the coffee break about the EU [European Union] or some other legal thing. Its certainly way, way down from three or four years ago, when we were in the height of some of the legal issues that got resolved.
So, really, by far its security. In fact, Longhorn, were thrilled about the innovation there; its our No. 2 priority but its been pushed back. We dont even know when it is, so we cant do precise math on this. But its certainly quite a bit pushed back because weve really prioritized the resources to be on security-focused issues.
With Longhorn being pushed back, what are the make-or-break products for Microsoft before then?
From an IT department, products like SMS, MOM, which is Microsoft Operations Manager, and Active Directory are kind of the infrastructure-level things that end users ideally dont even know that theyre there. They just know you turn on your system and its up to date. You log in from anywhere on the network and you get the right set of privileges. Those infrastructural things that require less manpower and let people deliver a higher degree of reliability in their systems and have this auditability to know whats going on.
For a lot of our customers, the basic dialogue is, hey, if your infrastructural products can save us cost, let us buy the industry-standard servers, let us use cheaper communications, reduce operational costs, simplify development costs. And particularly these infrastructural products, if they can do that, then fine, that will free up some of our budget to look at wireless and tablets and Web service applications and those things. But part of the bargain is that you, Microsoft, have to show us how technology is simplifying some existing problems in order for us to put onto our plate some of the productivity-oriented things that people know will come.
And so thats the typical dialogue, helping them look at their environment, say, OK do they have too much variety, do they have the right automatic tools, are they employing best practices and then translating that into reliability and cost-saving issues? And then typically, as we go through that, theyll say great, lets use that and deploy Office 2003. Lets use that and move up from file servers to Sharepoint servers that let collaboration be done in a new way. Lets try out and do a pilot of Live Meeting, do a pilot of tablet. And so our ideal customer is one whos got a mix of those innovative projects and those infrastructural reform projects going on in parallel.
Im really interested in modeling, and at your financial analysts meeting last summer you made the comment that modeling would be big in Microsofts tools. Im also pretty impressed with the Eclipse toolset and how theyre going about supporting modeling. Do you think that makes Eclipse more of a viable competitor to Visual Studio?
IBM has their Rational group with Rose, and they have some Eclipse stuff. So as is often typical, they have multiple, semi-overlapping approaches there.
Modeling is the future, so every company thats working on this I think its great, and I think there are some real contributions that can be made. You know UML [Unified Modeling Language] made the meta-models a little complex, so I dont think UML alone is the answer. Web services forces you to think modeling. And thats part of the good thing about it. And the promise here is that you write a lot less code, that you have a model of the business process. And you just look at that visually and say here is how I want to customize it.
So even a business could express in a formal, modeled way, not just scribbling on paper, how the business process is changing over time or how its different from other companies. So instead of having lots of code behind that, you just have visual, essentially model, customization.
So, I think we believe that. There are certainly some people from IBM who have that same vision, and I think itll be healthy competition between the two of us because todays modeling products fall short. Thats one part of Visual Studio 2005, that we do have some neat things coming along that will be part of it that we havent shown completely. Weve shown a lot of it, but Visual Studio 2005 is very broad.
So youre talking about Whitehorse?
Yeah, exactly. Thats the codename. But some of the Whitehorse stuff we havent shown publicly, like how it lets you visually design Web services and lets you visually design the relationship between the application and the deployment, this thing we call the Dynamic Systems Initiative, where the developer says, OK, what sort of resources are needed to run the application. And then the person who runs the application just binds that model to the actual execution environment. And then the developer can see if there are any performance issues or problems because they have this common model.
So, modeling is pretty magic stuff, whether its management problems or business customization problems or work-flow problems, visual modeling. Even the Office group now really gets that for document life-cycle rights management, that this visual modeling will be key to them. Business intelligence, where you let people navigate through things, is another area where modeling could be used. Its probably the biggest thing going on. And both Visual Studio and Office need to be on top of that. Theres a guy at Microsoft, Bill Baker, whos our business intelligence guy whos been promoting these ideas very successfully.
Be sure to add our eWEEK.com Windows news feed to your RSS newsreader or My Yahoo page: