Transcript of the Scott McNealy Keynote Session at JavaOne.

Delivered May 30, 1996 at Moscone Center, San Francisco, California.

Scott McNealy, Chairman of the Board, President and Chief Executive Officer, Sun Microsystems(tm), Inc.

Presentation entitled "We Are Already Hopelessly Behind"

With opening and closing statements by JavaOne host, John Gage, Director, Science Office, Sun Microsystems, Inc.

Gage (Host):

Good morning. I can see we're up early and we're filling the room, which I am amazed by. We're trying to keep track of what happens to a city when you loose 6,000 Java(tm) hackers in it. We have reports about parking meters being violated. There's a lot of this inserting component stuff going on, but we'll wait for a few days to assess the results of what you're doing. We expect to have every espresso machine in North Beach somehow made smart. Parking meters are a natural candidate. And anybody that can get into the cash registers at expensive restaurants should do something about automatic expense account filing, that's something that we'd be very interested in.

Now here's the guide for today. [Continues with details of the day's schedule.]

I would like all of you to observe my next item, which is the author ribbons. The authors are our heroes because they're writing everything we need to know about Java. When you see somebody with one of these author's ribbons, go up and say something to them like, "What a terrible chapter 3; there are seven bugs in it." I think we should do what Knuth used to do -- any author that has someone find a bug ought to give, I don't know, 25ยข and a mention or something. I think that would be good feedback from all of us about what we need to know about Java.

By the way, I was wrong again. It turns out the book list has gone from 97 to 108. There are 110 known English books that will be out by October. The list in Japanese has gone up considerably, and the French, Spanish, and German lists are going to be posted soon. So there's a lot of material out there for all of us to use and learn from. All of you who are authors, please come up and get your ribbon. I've handed out a number already. I signed them, and Scott signed them. In some iteration we'll get Gosling or someone else, who can actually read your books and grade you, to sign them.

I have a new candidate ribbon. It's going to be blood red. And it's going to be for all those people who ported the Java Virtual Machine to something, because they put in blood, sweat, and tears to do that. So I want those people to come up to me at some point during the next two days and identify themselves and we'll give them the blood red ribbon of honor. So those are my ribbons.

Now I want to tell you why it's so important to listen to what Scott McNealy has to say. When we began the Java project, as James described yesterday, it was a tiny "skunkworks" project. In any large engineering organization, the fight for budget is intense. All of the projects that are making money and show a return get funded, and all of them know that as year-end approaches, they'd better spend all of their money or it will be taken away. So every engineering organization is out figuring out ways to spend money.

And by using some of those funds, we established this small, secret, "skunkworks" project. It had to have a protector at the very top, and that protector was Scott. Because Scott said, "I don't care if what these guys do completely changes computing, if it changes everything we do now. It's important that it be done." So Scott protected the project, and it went on. When he first saw the "Star7," his instinct was to test the quality of this small device. How intuitive is this simple interface? So he handed it to his secretary. He didn't even try it first himself. There was no manual, the machine was the manual -- it was selfdocumenting. He asked her to figure it out. She did, and she could use it. It took her about two minutes to become completely familiar with what it did, which she discovered in the process of using it.

That's the criteria you've got to use in designing software that you're going to turn into a business, that you're going to turn into some engineering project used by people. Is it simple? Is it selfdocumenting? Can it be easily understood? If you can do that, you can create something of lasting value in the worldwide market.

So let me introduce to you, Scott McNealy, the President of Sun Microsystems.

McNealy:

Hello everyone. It's nice to be here and it's also pretty exciting. This must be a kind of a religious deal here. I can't imagine how we got all of you up here at 8:30. You must have had a cup of java or something this morning. Actually, I never thought I'd get up here on the podium. I didn't think John would ever sit down, but thank you for introducing me, John.

To warm up, I thought I would get it out of the way and do my Microsoft bashing right up front. I thought I'd do a top ten list of the reasons why people will keep their PCs. You know, we've got these reference profile network terminals coming down the road, and so there is no reason to have a PC on your desk, but all the industry analysts are saying they're definitely still going to be around. People are going to continue to want to have their PCs. So I thought I'd give you the reasons why you're still going to see PCs. These are not the reasons the analysts are promoting.

The number ten reason why people will keep their PCs is because it's the only way to see the Bettman Archives.

The number nine reason why people will want to keep their PCs is because charities won't accept them as donations.

The number eight reason, and this is a real reason to keep your PC, is because it says "Intel Inside" outside.

The number seven reason is because it makes a great 167-megahertz space heater.

The number six reason for keeping your PC is that it trains you for moonlighting as a system admin. (I didn't write all these.)

The number five reason is that you're still waiting for the depreciation cycle to end.

The number four reason, and this is my personal favorite, is because I want to add pictures and music to my spreadsheet. Very useful, very useful.

Very important here, number three, you can use the reboot time for a biobreak.

The number two reason: this is certainly true in our organization, is that it's the only CD player allowed in your office.

And the number one reason you'll keep your PC in your office, is because you love the white noise; it helps you think.

So there you have it. Don't worry, I'm not done yet. I have a few minutes to talk about the Java(tm) Platform. This is a different kind of audience to talk to, because it's a self-selecting group, so I don't think we need to get into the basics of what is Java is and why it's compelling. But, I thought I'd just give a quick overview. Eric showed a tornado. Twister is the new movie and we thought we'd have reusable concepts here. It really is changing. When you think that Java is a year old and Mosaic wasn't around three or four years ago. It's kind of funny for a lot of you. I like this group, actually -- it's kind of fun -- it reminds me of the "uniform" back 10 or 12 years ago. The only difference between the uniform back then and now, was that I was about the average age -- now I'm an old geezer compared to the audience. But, I think it's good to see the fresh, new, young energy and new blood coming after this like crazy. It's reminiscent of the old days, the excitement and the energy.

But today is a time when you just can't keep up. I don't think anybody could ever hope to go out proprietary anymore. We figured out a long time ago that the network was the computer. People come up to me and say, "Now did you really use that as a phrase in the 1980s? Didn't you just make that up after the Internet happened?" No, we actually had this. We had a group come up with that concept: the network is the computer. We trademarked it, then hired some more engineers, came up with some new phrases, mothballed it, and then we brought it back out, because it turned out to be the best one and the right one.

Also, we have always been "open." The two foundations of Sun are open and network computing. That hasn't changed. It won't change. It's what we built the company on. It doesn't seem to be hurting us on the stock market any more. In the old days, people used to say, "Why would you invest in Sun? They can't lock their customer in." It was literally hard to raise money. It was hard to get a stock price. If you were smart, two years ago you could have bought Sun for 30% of revenues at basically nine bucks.

The environment is a little different now, as people have decided that network computing does matter and "open" is actually the way to make money. You've heard all these proprietary trees falling every day. It started about ten years ago as open volume, UNIX, TCP/IP and all the other open interfaces moved through the computer industry, and destroyed every one of the proprietary business models. Until recently, there were really only two that were standing: Intel and Microsoft. Java is going to do a very interesting thing to those two environments in the sense that it's going to level those playing fields, make them open, and make them operate in open margins. At least, that's my guess. We'll have to see whether it comes true.

Looking backwards, everybody had to figure out what to write to. This was always the big challenge for us. If you wanted to write a desktop application, you wrote it to Wintel. But, you know, IBM has sold a lot of OS/2 out there. Then, there are people who want to do "real work" on UNIX. The Macintosh people have an incredibly strong and loyal following, and then you have this new thing called Windows NT and all the rest of it. What do you write to? It's been the big challenge of the industry, and that's why this group went off to do the Java thing. What does Java mean from my perspective and from the industry's perspective?

First of all, I believe it makes everybody compete. It makes the microprocessor gang compete, because, now every Java applet runs on every microprocessor the day it's born. You know, a six-month headstart with an application ported and compiled to your microprocessor gives you an incredible performance lead -- just sheer raw performance. Because in the six months it takes another microprocessor to get an application compiled to it, and in volume, you can move your megahertz, your performance, you can debug, get floating point to work, and those kinds of things, very quickly.

This is going to change the competitive nature in the microprocessor arena. When we launched UltraSPARC(tm), or when we launched SPARC(tm) in 1987, people said, "Oh my gosh, how are they going to establish a new microprocessor architecture? Is Alpha ever going to make it?" People are still asking those kinds of questions.

But, when we launched Java chips -- it's a new instruction set for microprocessor architecture -- not once did anybody say, "How are you going to make that standard? Are you really going to be..." because, all of a sudden Java, with its architecturally-neutral environment, says instruction sets don't matter. Price, performance, quality, accuracy, service support, and availability; those are the issues that are going to make microprocessors win; process technology, and those sorts of things. All of a sudden, instruction sets don't matter. That's good news. It's good news for those companies who've been making money from microprocessors and not being Intel, because now when we have access to the total available market, as opposed to a very small percentage of the market, things will go a lot better.

The second thing that's happened -- and this is important from Sun's perspective -- I grew up in Detroit so I always think of automobile analogies -- we are in the light truck sports utility vehicle kind of business with our workstations and servers. They're very high margin, not incredibly high volume, and it's hard to support an automobile company if you're just doing vans and pick-up trucks. You need the volume of the passenger car market. We didn't really have a passenger car business. That was locked up and owned by the Wintel space. Now, with Java clients, we have a way of getting into that passenger car unit volume. In fact, it's much greater than just passenger cars; it gets us into motorcycles, mopeds, and bicycles, and the whole deal with Java clients all over the map, starting with Java chips and very low-end machines.

So this is a big, big breakthrough. We were trying to solve different problems with Java besides just making Sun successful. The nice thing about this team was that they were told to go off and do something for the customer -- something neat and exciting. These are the things, at least in 1996, that we are trying to solve. We're trying to solve the desktop-enterprise computing problem. You'll see people do kiosks in the retail environment and you'll see people bundling them into telephones, talking about the home consumer, and all the rest of it. There's already been some negative feedback. For example, "Oh my gosh, there's no way this is going to replace the PC," or "There's no way this is going to get up in the home," or whatever. Java is meant to solve some of these very difficult problems.

You've all heard me rant and rave about what a challenge it is to put a Wintel computer, a gigabyte mass-storage disk-farm file system with 32-megabyte, 32-bit multitasking, multithreaded SMP environment, with all the bells, and whistles, and middleware, and a backup medium known as a floppy, a software-distribution medium, or music distribution medium known as a CD player -- put all this on a desktop. There are no two PCs on the planet that are the same. They're like fingerprints. There's absolutely no two that are the same. Even if they have the same software when you boot them up, they'll come up differently. And by the time you get through configuring and organizing them, and every PC user is probably a better system administrator than most VAX/VMS administrators were ten years ago. They get more practice at it. The VAX would go down once every couple weeks, right?

The best way to get press on this network terminal is to call it a $500 terminal. The least interesting, least compelling aspect of this network client or the Java client is its cost. The zero administration client aspect, or the zero administration focus of this thing is the number one issue. Once you start to deploy it, security becomes a big issue, and for those of you in the enterprise world, and for all of you authors out there, I'll spend a little more time talking about software distribution advantages. The compelling pieces of the software distribution components of this environment are just outstanding. They change the nature of who's in charge in the software business.

These are the problems we're trying to solve in 1996. We're not trying to solve home computing yet. We can do that next year. We're not trying to replace the PCs absolutely. There's plenty of time for that. I've always had this theory that takes as long for something to go away as it took to get there. We've been doing PCs for quite a while. They'll be around, unfortunately -- I mean, we haven't solved the common cold, either.

What Java means to me is that we now have a common language that we can all speak. I might as well just continue to pick on French, right. It's the written and spoken dialect that I learned, so I always say it is my second language -- though it's more like my fifth language in terms of how well I do with it. But in the written and spoken language world, there would not be French if there wasn't a government mandating that it be spoken, right? Well, there's Quebec, I guess. But the point is that we go back 10, 15 years ago, we had dozens of written and spoken languages, and we had so many dialects. Remember Wang OS? Remember VMS? If you go back through all the different languages -- because there were no governments mandating that somebody speak a particular language --we had a very Darwinian situation with respect to the written and spoken languages of computers.

Nobody owns Japanese; nobody owns English; nobody owns French; but you have to pay a right-to-use license to use the two alphabet characters, N and T. There's something wrong with that, there's something very wrong. You had a situation where groups of people were not speaking the same language, so you had to fundamentally innovate on a dozen different platforms if you wanted to get unit volume around the world.

I think this is an interesting opportunity -- I mean, how many of you software developers raises your hand when the boss says, "I need somebody to re-port this?" Is that exciting? Is that really how you wanted to use your education? Re-porting? That's like being in the army you go dig a hole, fill it, and then dig it again You shouldn't be here if you like to port.

What we're saying is that with the interpreted, secure environment, porting goes the way of the punch card. It's not the way to do it anymore. I was with a little customer in Bentonville, Arkansas yesterday. You might know who that is, I wasn't meeting with the governor. Do they have one? (I digress.) I apologized to the customer; I had been there about nine months ago when we explained to them the Java concept and how it could help in the retail marketplace. They looked at me as if I was on drugs, and thinking, "Where did this guy come from?"

I began the next meeting with, "I apologize for totally underhyping what was going on with Java nine months ago." Who would have believed nine months ago that the entire Windows suite of -- I don't know how many different operating systems there are at Microsoft; and all the IBM operating systems; the Apple environment; all the UNIXs except DEC I think -- does that matter? -- Novell; and now all these Java clients; and the reference profile NCs would all have the exact same embedded API, plus the browsers; and then there's the JavaOS -- running on bare metal on any microprocessor you want to port it to.

If I had said that nine months ago, you'd have thrown things at me, but it has happened in very short order. It's something we certainly didn't predict, and I would hate to say I'm a good visionary; if I was I would have predicted nine months ago that this would happen and this would be the situation. But I'll tell you that this is very unusual -- it's unique. This is the first time we've ever had an API that cut across all of these platforms.

I'm a little worried now about Windows. I fear that it doesn't have enough critical mass in terms of unit volume. I hope they survive. We need competition. It's kind of interesting. I think it's the first time Microsoft ever licensed software from a company without buying the company too.

I think it's also interesting that people are committing to staying compliant, committing to giving back enhancements and strange proprietary directions. They're committing to staying compliant within x number of months, so that we can move this standard forward. We've been in the UNIX world. We've been there and done that, and seen the fragmentation. I think the fragmentation, to some level, is important in the innovation. What we have with Java is a process that allows innovation but minimizes the fragmentation and brings it together on a very regular basis. This means that you have a constantly evolving API, without the fragmentation, so that you get the unit volume of all of these operating environments and all the network clients. I think we've learned a little bit from the UNIX space.

This is exciting. This is a big change in the industry and it will be interesting to see how it all plays out. I believe it's going to win first in the enterprise, then we'll get into all those other pieces. Now you've seen a few announcements of places where all this is going to happen, and we've been seeing every hub, router, switch, printer, copier, set-top box, game machine, desktop computer, nomadic computer, server, telephone handset, cellular phone and automobiles -- all of those environments are going to be zero admin clients. Because if you think about it, all you need to create a Java client, or a reference profile computer is a network port and a microprocessor. If you have those two things, you can turn anything you want into a Java machine.

You've all heard about Nortel. [Speaker holds up small cellular phone] This one happens to be Motorola, but this is a perfect opportunity for a network computer. There are 300,000 lines of code already in this little machine, and, by the way, it's smaller than that, because that's the battery! If you look at this thing -- it's got a keyboard, a display, it's got a network port, it's got a speaker, it's got a microphone, it's got a microprocessor, and it's got storage in there. This is the opportunity. And if I say something really silly, you can be in the audience and type in [speaker spells out] "s e l l s u n w" on your Java applet and you can beat everybody to the phone.

So this is the kind of environment that you will see out there in the marketplace. It's real, it's happening. The fact that Nortel has already announced that they are putting the Java environment in their telephone handsets is part of the vision that we have had in terms of making this operating environment ubiquitous. Ubiquitous everywhere -- isn't that an overstatement.

I'd like to talk about the desktop environment, which is one of my favorites, and it's certainly popular with the press. Think about what you have on your desktop. What do you really need versus what do you want? I know you want a disk drive on your desktop that's not controlled by the MIS department, or the telephone company, or by an Internet service provider, because there's stuff you want to, like, download, pictures from the Net that you want to keep local, right? I actually get a much higher percentage of laughing out of this crowd than most when I say that line. You might want to keep your resume or some other things. People still stuff money in their mattresses. People think data is like money and they want to keep it local.

Actually, having a disk with you is a very inappropriate and inconvenient way to handle your disks. Your little disk is very expensive because they've got to make it small. You run out of space all the time. It never crashes, you know that, right. Nobody ever spills a Coke. You never drop it, and I'm sure you back it up everyday. You can see the problems with that environment as opposed to having your disk drive out over the network, managed by an organization that knows how to run a disk farm, that has mirroring, and uninterruptable power supplies, and has very high speed, low cost disks that are much more efficient. You never run out of space, and somebody gets fired if they don't back it up everyday for you. That's the model for keeping your mass storage. And it's so much more reasonable.

And so our vision is that someday people will go to their PC; and they'll take the CD out and they'll give themselves more DRAM. They'll take the Intel chip out and put a fast low-cost chip in, like an UltraSPARC(tm) now, there's a vision -- and, by the way, why don't you throw two or three extra ones in there; take your floppy out, and give yourself a faster network port. You don't have to do user administration so you can afford to move -- I don't know where we are today with 50 users per system administrator -- and get to a thousand users per system administrator, like a telephone model.

Most importantly, you can take your operating system and get rid of your disk if you like, or you can leave it there as a local network cache, but erase it. Erase it. Get rid of it. You don't want any data there. It's just a disk cache for downloading applets if you want, or get rid of it because you don't like white noise. And just download the JavaOS and run the JavaOS on that environment. That's the model. In fact, I am suggesting to a lot of our customers -- they're all trying to figure out now, "Do we go to Windows95 or do we go whole hog and really mess with the productivity of our employees and give them WindowsNT?" I look at them and the first question I ask them is, "What's your second sourcing strategy for WindowsNT?" and they all look at me like, "Oh really? Second sourcing. That's important?"

But why would you give somebody WindowsNT? That's not an upgrade. In fact, I'm asking customers to think about downgrading. Take your Compaq computer, throw out the CDs, throw out the floppy, erase the disk, and put the JavaOS on it. I call it downgrading your PC. And it's a much cheaper way to move forward into the next century by downgrading your PC to the JavaOS as opposed to upgrading it to WindowsNT. In fact, you can't upgrade it. You've got to go buy a new one. And this is really the opportunity. I've asked the JavaSoft(tm) folks to figure out how to create a floppy, or something, that can be downloaded through some sort of bootable Ethernet port with JavaOS on it for all PCs out there. So that all the CIOs out there who want to move their world from the PC to the Java client, will have a way of doing that. Now I don't know whether people will actually do that. I'm partially skeptical that people will ever do that, because taking a PC away from somebody would be about as easy as taking a mountain bike away from somebody. They're so wrapped into it and fired up about the environment that I think it will be very difficult to do, but it certainly will be fun to go in and offer that to the CIOs.

If you do get people to move, what are the top five uses for PCs in the year 2021? Number five reason is lava lamps. Number four is fish tanks. Number three is that it replaces grandpa's "I used to drive a Studebaker" story. For those of you from the Big Apple, it will fill the New York City potholes. And for everybody who uses the highways, they could be used for recycled scrap metal in highway crash barriers. So we do have something to do with all those PCs when we get done with them.

What I'd like to do now is talk about some of the more recent announcements that we've had. Actually, this one is so recent it's not going to happen for another three hours. Sun Microelectronics(tm) announced that Nortel is going to put the Java environment and Java chips into their telephone handsets. Now, Nortel is a big, wonderfully successful company, and we're very excited about having them as a partner. But we have four very, very large global consumer electronic, and technology companies that we'll be announcing at noon today that will make Nortel look small. We encourage everybody to stay tuned for that. The press knows where to go for that one. Put that together with Nortel and I think Sun Microelectronics is on the way. We're pretty excited about it. So stay tuned for that.

You might have picked up the Mercury News from San Jose. Corel is doing the WordPerfect Office Suite in Java. Cool. This is going to be your first object-oriented extensible, non-ported office environment available to everyone, and we really think this will drive... well, we'll see what happens with it. I don't want to overstate, but we're pretty excited about it. And we're pretty excited about running this product on the Java client that we'll be announcing and shipping in volume this calendar year. We've said that before, and I'm saying it again here, that we're still on track and rumor has it that some customers are already starting to see some early versions out in the marketplace. That's just a rumor.

Java has gotten so big that governments are now adopting, signing up and licensing Java. The Taiwanese government is very, very excited about putting some competition and some choice into their manufacturing operations. They have already licensed Java and are working with 22 different manufacturers in Taiwan to get them set up to build Java-based clients. All of a sudden we now have one of the most impressive global manufacturing engines in the country of Taiwan, 100% behind this whole architecture. So again, it's just moving. Maybe this one will be around awhile, because now even governments are mandating this written and spoken language of computers. So again very, very exciting.

Now I'd like to finish the keynote and let you all get on to the slightly more technical aspects of this event. We'd like to make some suggestions for you all as you move forward, and I know every time I go to a customer, they ask, "Well, when should we start thinking about this and how do we go about it? How do we get Javatized?" Let me make some suggestions.

The first suggestion is -- get going -- get on network time. Network the dog knows all about that. A dog lives seven years every year, and I think that's the situation we're all in now. I don't know how to say it more aggressively other than tell you about my visit to the CIO of a trucking company in Jacksonville, Florida. The CIO said that he had edicted that any project inside of his organization had to -- from the time it was authorized to the time it was in production -- be done in six weeks or less, or the project was not approved. Fundamentally, this CIO was edicting Internet time or network time. And I think we all need to start thinking about that and looking at that. In fact, the Java Environment with the small, object-oriented scalable capabilities, the simple environment that it is, allows you to actually do six week projects and re-architect the way you do your development.

The next suggestion -- this is not new, Gilder said it -- don't bet against bandwidth. The networks are going to get faster and faster. And they'll get faster more quickly than you'll be able to respond to them. There's just no question in my mind. It's all going to happen. I was in Boston -- this picture depicts T3s to your barn, or something, I don't know -- I was in Boston earlier this week at the Internet Conference there, and somebody was telling me that there's an apartment house in downtown Boston advertising that every apartment in the apartment house had a T1 line dedicated to it, and that was actually a feature of the apartment building. Cool, huh? I still wouldn't move there. But you need to start thinking about what it means when that kind of thing happens, when the cable modem happens, when ISDN starts happening, when 100 megabit and ATM, the new wireless environments, and all these other things -- when these things start happening, it definitely changes things quite a bit.

Another suggestion is "think easyware." Easyware is -- I go back ten years ago when we were all worried about GUIs and trying to come up with ways that people could become familiar and comfortable dealing with the arcane, and absolutely nonintuitive ways of doing things on screen. The new way to do it is just not to have a GUI, because if you can read it or if you can see it, you know how to work it. Building easyware is all about making the user interface just "brain-dead obvious." Like an on-switch or something like that. And this is the opportunity we have, to eliminate the concept of a GUI, a formalized GUI, and get to the concept of doing something that is just "user usable."

The second component of that is building simple applications -- I often get asked the question, "When am I going to see a sophisticated application in Java?" And my initial reaction is almost: never, I hope. Because the whole concept behind this thing is to build easy, simple applications. It's got to be sophisticated in its simplicity, and sophisticated in its ease of use; and that's the real opportunity we have with the new application environment, to not build a sophisticated environment. My hope is that we can do that.

I'd like to turn it over now to Miko Matsumura. Miko is going to show you some of the applications being built that are very sophisticated, but still quite easy to use.

Matsumura:

Howdy. Thanks, Scott. I'm the Java evangelist for JavaSoft. And I'm really pleased to be here talking to developers. I just wanted to paint the screen with some pixels that were actually Java-driven, add a little bit of excitement to your lives and to my own life, because this is what they call a live Internet demo in front of what, 5,000 people? Okay. So if we could just get this screen up onto here, we can start to see some of the things that I'm working on.

As an evangelist, I travel around the world describing Java. And it's an extremely interesting process because of what it is. Well, starting out, it was a programming language. People got into the JDK, and, all of a sudden, it was a browser.

[Miko launches a demo in Netscape Navigator 2.0.1]

Now what I'm going to show you is something about easyware -- about componentware. It's very difficult to describe, but I hope you can just watch and see what kinds of things are possible. So for example, this is a Web page. We're looking at Netscape Navigator and I've just opened up and drawn a text field onto Netscape Navigator. Pretty fun stuff. This is an applet, and I can resize this text field and you've got this -- and it's movable, which is a lot of fun.

[You can try this at http://www.eas.asu.edu/~lu/VJ2/example1.html]

Now notice what I'm going to do over here. I'm going to do something that's pretty classic, and it's from the history of Steve Jobs. I'm going to draw a scroll bar; in fact, I'm going to draw a really, really big scroll bar so that you all can see it fairly well. [Miko draws on the screen] Now what I can do, if you really want to have fun... On the panel to the left, I'm connecting the scroll bar to the slider, and as you can see, the numbers are changing -- very Steve Jobs. So far, so good... this is fun... Of course, you can draw a button, and I'm going to draw a really big button... there's a big button... great. And that button of course is a component functionality.

So what is this? People think, "Okay, this is a GUI builder." Okay, GUI builder is fine, but this has moved from a language to a browser. Do you remember the confusion people had when they said, "Well, what is Java? Is Java a browser?" People really got confused. It's not a browser; it's not a language. Now people are getting confused because they're thinking okay, Java is an operating system. So what is it? We're having difficulty describing it. And the reason we're having such difficulty describing it is because fundamentally, it's a revolution. It changes the way we think, changes the language, and it makes it extremely difficult for people like me to explain it. So resort to doing a lot of demos, which is really interesting.

What I want to show you now is -- if you look at this stuff on the left-hand side, [Indicates the blue window on the left] it appears not to be a rapid application development environment, but to be like an operating system. What I'm doing is connecting those components you see on the left with pipes. So I'm actually piping instructions between executables. Does that sound like UNIX? Very interesting stuff.

So now we have a dog, right? [an animation of a running dog appears on the page] And this running dog is Network, and he's running seven years to every one of our years. What I'm going to do is take this button, connect it here, and hit the button. [Miko clicks the button] Now the dog is going to run backwards. It's pretty fun, huh? This is where the components start to get a little bit more sophisticated than just GUI components. I can do that little Purina Cat Chow thing with the button, you know. [Clicking the button repeatedly, he causes the dog to run backwards and forwards.] That's pretty fun. And just to show you that this component has reusability, I'm going to take the scroll bar, disconnect it from there, and connect it to the speed thing. Notice now he's going really slowly when I scroll down here, and now he's going really fast when I scroll up here. It's pretty fun. So I can make him go backwards, really slowly, really fast. Neat stuff.

The point is that these components reach a fairly high level of sophistication. This is a graph, so you can have a pie chart or a bar chart. There's a spreadsheet component. There's an SQL component. All these components can be interrelated visually, and remember I'm actually drawing this on the Web page. Think about that. Pretty amazing. This is a research project that's been done by a professor at Arizona State University. Very interesting stuff.

[Miko launches a demo of the ESRI-mapping software]

[Note: ESRI-map demo not available on-line]

Where does this lead? Well, it leads to easyware. And easyware is essentially the kind of thing you want to do to make things easy for people to get relevant information. Now, we're going to go to 747 Howard Street, the site of the Moscone Center. And I'm going to show you a demonstration of a mapping front end. By the way, that last demo was completely Java. In this demo, the front end is completely Java, in fact, it's an applet running in Navigator. What we're seeing is a coffee cup at the Moscone Center address which is right here. [The screen shows a detailed map of San Francisco] If this were a Microsoft demo it would be called "Where do you want to eat today?" What I'm going to do is search, I can search by point. So I can say restaurants. So I want to find a restaurant here. I'm going to search in this big circle. I'm going to say, "Where in this big circle can I eat?" Boom. Restaurants pop up. I can say "identify." Let's identify a restaurant here. That one says it's Maxfield's American. Over here you have a Chinese restaurant. You can open up a table and here are all these different restaurants.

You can click here, [he clicks on a restaurant icon] and if you really want some fun, this is actually linked to a Home Page for that restaurant. The Home Page gives you a rating, gives you the times, gives you a phone number, and you can make your reservations right there. Pretty cool, pretty relevant, pretty easy to navigate, I would say. Also, if you want to, you can say , "I need an ATM machine. Why don't I search?" This time I want to search in a box around the area of that restaurant -- and look, there's one right there. So you can use that ATM machine.

But this is my favorite feature -- a layer that I added yesterday called the Nerd Layer. So I'm a nerd and I'm hanging out and I'm wondering where I want to go and be a nerd? So I'll search by making a box. Let's make a big box here and look for nerd sites. Here's a couple right here. [Little icons that represent "nerds" pop up on the screen] Little Home Page button icons. Pull up a table, and there's one called "The Box," which is a club, and "Hot Wired," and there's a site called "Aereal," and you can click on any of these. Let's click on Hot Wired. It will pull up the Hot Wired Home Page. You can find out where to go. You can find out where to eat. You can find out where to bank. Basically, you can navigate around. Imagine that kind of thing happening in a handheld wireless browser environment. Now you're really cooking with gas.

I guess the question of the day is: "What is Java? Is it a language? Is it a browser? Is it an OS?" Difficult to describe. Fundamentally I think it's a revolution and it's a revolution that's going to lead to componentware, easyware, and it's going to add a lot of relevance to a lot of people's lives. So thanks.

McNealy:

Thanks, Miko. Can you track down the nearest McDonald's for me at lunch here? I'm ready to go. So I'm going to try and finish up here with the rest of the suggestions. Obviously, what Miko showed is that you really should start writing your front-ends in Java now, and as you move forward. It's time to start rewriting the legacy apps in Java. The development tools are there. Everybody is targeting their development environments. We would certainly like you to check out Java WorkShop, but there are all kinds of Java-based development environments that you ought to be looking at, even Microsoft is targeting their development environments for Java.

Prepare for the wireless world. Bandwidth is happening, and wireless is happening -- you're going to see wireless connectors happening everywhere. I saw that a consumer electronics company now has -- do you remember the old Motorola brick phones? A little smaller than that -- a cellular phone that opens up like a clamshell; has a Sharp-like organizer inside of it, with a QWERTY keyboard, and it can do fax, and email and all the rest of it, all in your own personal communicator. And there are people talking to us about putting Java chips and the Java Virtual Machine in that environment. This stuff is going to happen, and it's going to happen very quickly, and you need to think about your applications getting into these nomadic and wireless devices. You need to architect them in a different way.

You're not going to be able to download Microsoft Office as a Java applet. You're going to have to reshave between the time you hit download and when it shows up. So you really do need to think and architect your applications a little bit differently.

Obviously a very appropriate suggestion here, and something that I think Java allows you to do, is that you need to understand and adhere to the security features. You know nothing is entirely secure. If it was totally secure, you might actually have a situation where you could never get at it again. But if you pay attention to the security features, you can build the most secure environment of any environment that we know of today. And more importantly, by using Java you can just say "no" to memory leaks. I have been to several accounts recently. One company said they shut down the cellular phone service in a particularly large geography on a Friday afternoon due to a memory leak. There's another major Internet service provider who was basically being held up in terms of the number of people they could sign up because of a memory leak. We threw more microprocessor at it, and it just leaked faster, right?

The challenge and the opportunity is to move to an environment where memory management and garbage collection is done. And if you look at it, as much as half, or more, of all of the bugs that you face out there, and the lead time, and the deployment time of applications is with memory leaks. So that's a big opportunity.

The next suggestion I'd like to make to you is old UNIX history coming out. When we did UNIX, we all started with the same code base; the same environment; and it was all compatible; and you could actually move from one environment to the other. But then we as, for-profit CEOs, and engineering managers, and all the rest of it hung a proprietary hook out there for you to go after. Some extension to UNIX, something outside of spec 1170 -- you know, spec 1170, one P for Proprietary. It was unique. And we dangled it in front of you software developers or in front of the ISVs, and said, "This will make your application even better. Just go for it." And we even, you may have heard me say, we even like stuck a Twinkie on the end of it, because there's no programmer in the world who would turn down a Twinkie. And boom. You went for it, and you made your application unique to that version of UNIX.

It's happening big time again on the Internet, and it is happening in the browser world. It used to be anytime you put an HTML page up, it was viewable on any browser. Now all of a sudden, I'm sure you've all seen the little phrase, "Looks best under blah, blah, blah." Or "Only runs using blah, blah, blah." In fact, Microsoft is running around bragging about how many pages now only run under Explorer out there on the Net. They're actually cataloguing it. It's like they feel like they bagged it -- it's a big game trophy every time they get a page that only works with their stuff. That's not what the whole network is all about.

So my suggestion, and if you remember any of them, remember this one. Before you release any content to the Net, whether it be a page or an applet, or a video or whatever, test it under multiple browsers. Get Navigator. Get Explorer. Get Oracle, get whatever browsers you can find; Spyglass or whatever. And the HotJava(tm) user interface. The reason why I say HotJava is because today HotJava is the only browser environment that is all written in Java so that you can actually run it on the Java clients. I mean, you can't run most of the browsers out there without an operating system today. Java you can run without an operating system just by running it on the Java Virtual Machine on bare metal, the JavaOS.

And so if you test it on all of those environments and make sure it runs at speed and equally well in all of those environments, you have obviously used standards that have been accepted by everyone. If it looks better under one environment, you're getting yourself down into the proprietary trap.

Now in the UNIX world, it was shame on the vendors for providing those proprietary hooks, because it was a new deal, we faked you out. In the Internet, it's shame on you, if you don't make sure your stuff runs on multiple platforms, because you've been warned, you've been there, you've done that, and it's an opportunity now to force all the vendors to stay compatible, stay open, and stay multivendor. You have to do it as the content providers to the market. So take that into consideration.

The next thing is to understand that we're in a new model. The Java and the Mosaic model, right? Create the product -- early days -- you don't have to have it totally done. Launch it out on the Internet for free. Gain ubiquity, brand, and then start charging a little bit for upgrades. That's the model here. It's very different from the "go out, spend a gazillion dollars engineering release Windows 1.0, it doesn't work; do 1.1, it doesn't work; release 2.0, it doesn't work; 2.1; 3.0 sort of almost works; and then launch 3.1 with the Rolling Stones or something." Right? That's the old model. And it's a little difficult.

The first time we ever tried this model was in 1983 when Bill Joy came in and said, "We've invented NFS." And we said, "Great, how much can we charge for it?" And he said, "Let's make it open and available to anybody." We looked at him thinking, "You're nuts." He is, but he's a great guy. And in two weeks, he had us all talked into it. And no, we didn't make a lot of money on NFS, but we made a lot of money selling NFS-based computers, where we'd have been locked out if there hadn't been NFS. That's the model, and that's the big opportunity. The model works. You've got to trust it; it's like leaping blindfolded, but you just have to trust it. If the market won't take it for free, they certainly won't take it for money. So what have you got to lose?

The next thing to understand is that the Internet has changed the economics of software distribution, publishing and creation, because never before in any environment have we had a situation where you could write an application, publish it for free with zero cost to goods sold, like Marc Andreessen. Everybody says, "Oh brand is so key." Are you kidding? How much money did Andreessen spend on brand for Mosaic when he was at the University of Illinois? I mean, we spent more money today on Java than we have in the past twelve months in terms of promoting it. The brand. Java is a bigger brand than Sun now. And you haven't seen any TV ads for it. Brand doesn't matter. What matters is great authors. And this actually empowers the people in this room in a big time way, because in the old days, the publishers had all the power. The Microsofts were great publishers, lousy authors. That's a fact. But very, very powerful publishers and therefore very successful.

This new model says the powerful authors are in charge here, not the publishers. And this is a big advantage and a big opportunity for the people in this room. And I think you need to think about that big huge change. The Internet is a beautiful distribution mechanism. Everybody on the Internet is a computer user. How more targeted? If the catalog companies could have that kind of absolute targeting, you wouldn't get all that junk mail. This is really a very targeted mechanism for distributing your product.

And then finally, you are in absolutely no danger of moving too fast. Bill Joy came into our staff meeting and hollered at us and I looked at him, and I said, "You're absolutely right." This stuff is moving so much faster than we realize. And if your organization is more than five people and more than a few months old, you're hopelessly behind. That doesn't mean you're going to go away. It just means you're going to miss all kinds of low hanging watermelon out there in the marketplace, and you're just going to have to deal with that. But you can grow and be very, very successful, if you just decide to go like crazy in this direction. It's going to be hard to beat CA in mainframe software. It's going to be hard to beat Microsoft in doing ActiveX OLE stuff. But you have a chance here with an open, level playing field, with a big market that has a superset, a huge superset in terms of unit volume of the Windows ActiveX world to go like crazy and to establish a position like you have no other opportunity, I believe, in the computer business today.

So it's a big opportunity. It's a big chance to go make money and I think there is money there. You've got to be willing to grind it out and you've got to build great product. The interesting thing about this environment is it isn't going to be your brand; it isn't going to be your ad campaign; it isn't going to be the size of your organization; it isn't going to be who you know -- it's going to be how good your product is. And it's going to be how well you listen to the customer; how easily you made the product; how extensible; how open; how you made your product run on all browsers and all network clients; and gave yourself a total available market. Then it's going to be word of mouth. You put that product out on the Net, and it's going to be word of mouth that drives it and makes it happen. The interesting thing about word of mouth is that it's usually the best reference you can get on your product. It's way better than a cover story. It's way better than an ad campaign. It's way better than any big conference. It's word of mouth that will drive your product and make you all successful, and generate some money for you.

So I wish you all the best of luck. Sun is here to help. We want to help you. We want to help leverage your ideas and your activities and your energy. So make sure you get a hold of us. Use us, abuse us, and let us know what we can do better to provide a better opportunity for you all to make some money out there on the Internet. Thank you very much.

Gage (Host):

Well, remember you can reach Scott at scott.mcnealy@Sun.com. For those of you in the press that want to reach Scott McNealy, he's going out the door and into Room 300, just outside this room to the right, for a press conference, and as Scott says, he would not want to overstate anything, so he's going to be a mild and understated person.

I have a prediction to make. You've seen the money we've spent on this, and as Scott mentioned we've spent more today than we have in the history of Java, or in the history of the Grain project, or in the history of the Oak project, just to do this conference. And it's very pleasant to have this.

My prediction is -- given the announcements at noon, given the Nortel announcement -- that the next time we do this, which may be shorter than a year from now, you're not going to get any backpacks, and you're not going to get any bags with Sun logos. You're not going to get that expensive Java jacket that Scott's got on. You're going to get one of these little devices, because the cost of these is going to plummet, and you're going to be able to sit in the audience, and you're going to be able to "sell Sun W," or whatever other company, right there.

So today when you read in the Wall Street Journal that entire spread about the Web, what's going on, and Java-based things springing up all over, a year from now we'll give you -- it's my prediction, and you can hold me to this, and Scott. I think this is going to happen -- we'll give you a device in your hand that has these components assembled, not only the hardware, but the software that lets you read and write in a new way, and that's the metaphor I think you have to carry with you for the rest of this day and tomorrow. Find the powerful metaphor that lets you create something that can be used without explanation.

So with that we're off. Credits -- Tony Faustini is the professor at Arizona State University that did that componentware thing you saw. It's elegant. Every couple of months he adds another component. That's worth investigating. The geographic demo that Miko did is the ESRI demo front-end Java working against a serious database on the back end. So it can be done. It's being done everywhere. Go out and do it. Thanks a lot.


Scott McNealy/JavaOne Page