Michael McGuire:
Michael McGuire, Bloomberg Business News. First of all, I am very glad that Skottish is your language of choice, but apart from that, last night in New York your shares went to a five month high, or so I think on speculation, that you are going to introduce your first line of network computers by the end of the month. Is that correct and overall, how do you Š what do you see these network computers doing for you, bottom line?
Scott McNealy:
First of all, we don't comment on rumour - no matter how accurate or silly it may be. That's kind of a party line and it worked very well during the Apple rumours. If we had told you all what was really happening last spring, you would have stopped writing about it and we found it was an immense way to get a lot of press. So I can't comment on any future product announcement. We have promised that we would ship the Java Client environment before the end of the calendar year - we're on track to do that and in fact, pretty excited about it Š our customers are pretty excited about it. The whole concept of this environment is to basically give you a desktop computing environment that is as easy to use as the telephone. You don't think about administering your telephone. You don't think about loading software on to your telephone. You don't think about administering it or backing it up. We will have that kind of an environment for you - and it's not the desktop that matters - it's the system. It's the servers, the software, the network administration, and all of the other components that allow you to operate in what we call the zero administration Š the zero client administration, kind of environment. Its computing for people who want to compute, as opposed to computing for people who want to Š I mean we have some engineers who buy their ten speed bicycles and pay extra to have them unassembled. The network computer is not for those types. Its for those who have things to do and want to get things done and it is going to be an environment where you can just click and download your applications, surf the net, store your data in a databank, where it belongs, and we think that is a very powerful concept and one that is going to, not replace, the PC, I mean we haven't been able to solve the common cold either, but it will definitely be an alternative to Š you know, we're still doing mainframes - we should have stopped doing those a long time ago Š but it will be a long time before we eliminate the whole concept of a personal computer. It doesn't make sense but it will continue on. We think this gives people an opportunity Š we will be able to put kiosks in every retail environment, every car dealership, every clothing store. Can you imagine putting and NT kiosk in every clothing store? You would need a system administrator right there in case somebody actually used it. It's just not feasible in that kind of an environment, so I think you are going to see if we never replace another PC, you're going to see these Java Clients everywhere and coming from lots of different companies, you are going to see it embedded in your stereo equipment, in your Š oh, we were talking to a whitegoods company, and why a Java chip and the Java virtual machine couldn't be installed in your washer or dryer. It's not ridiculous. The cost and the environment is very, very, low, from a cost perspective and you can start embedding intelligence, networked intelligence, into all of your appliances.
LAURIE WILSON:
If you do have a question, I'd just like to say please indicate to either our floor manager - who's the lady crouching in the middle of the floor with the interesting headgear there - or Leigh Catlin - who is the gentlemen standing at the front. Our next question from the table in front.
Julie Rowbotham:
Julie Rowbotham from the Sydney Morning Herald. Taking up your point about the $150 a year tax that we're paying to Microsoft, I am wondering whether there is a risk that we'll swap that for a tax that we pay to the Telcos. How are we going to pay for our networks in this new environment?
Scott McNealy:
Well, we are going to have to pay the Telcos for networking and for communication, whether you have a PC or a Java Client, because communication, we are willing to pay for. I mean, we are just very social people and we need to communicate, so I am not sure that that is going to Š that the network client is going to have any impact, one way or the other, on that environment. I believe it will make the computing environment a lot less costly, because today if you want a five-feature word processor, you got to go out and buy Word. You get all 4,000 features - whether you want them or not. And you have got to pay for it all up front - the full cost of it. And you don't get any upgrades with it. In the new model, you can use the network, download a very tiny little Java word processing aplet, for free. Why for free? Because the author is absolutely willing to put it out on the net for free so that you can use it whenever you want, because an advertiser, like MacDonald's, is paying them to put golden arches around the border and have a hamburger button on it. Right. And whenever you need a word processor, you get the latest, bug-fixed version, and you just, fundamentally, get it for free. That's a lot cheaper and you can then pay for extra networking costs in that environment by not having to go out and buy these super-set, application packages that exist out in the marketplace today. So, I think there's trade offs, both ways, but I think the communication opportunity is a net win.
Tony Simon:
Tony Simon from the Australian, newspaper. You talk about the lean and mean Java applications and the bloated Windows 95 and Office, but I find that when I first used Netscape, Netscape browser, it was about one megabyte to download, then it grew to five megabytes, now Netscape, with all the plug-ins, with shockwave, real audio, and so on, is virtually as big as Windows 95 itself. How does that improve things, by moving to a Java-based world?
Scott McNealy:
You should be asking Jim Barksdale that question not me. That's a good question you ought to ask him. We will be introducing a Java-based browser environment. It's called 'Hot Java'. We have already launched it and the Hot Java is a very small, very efficient, very capable, Java browser that does HTML and runs Java applications. That's what you want. That's what you need. It is interesting to watch the thermo-nuclear arms race between Explorer and Navigator to see who can create the biggest hairball. I think that's pretty interesting. I am not sure that in the dedicated use and corporate environments, that we're going to be selling these Java clients to, that you want all those features, that you want all those plug-ins, that you want all those Microsoft applications. You don't need them, you don't want them, and in fact you don't want to give them to users. A classic example for a Java client would be a bank teller - where you run like six or seven applications, that's it - you don't do Word or Excel. I mean its just Š there are some very clear bank-created, MIS-department-created aplets that you run and they are downloaded at night when the network is not being used anyhow, when rates are cheap and your eight aplets show up every day with the latest interest rates, the latest exchange rates, the latest stuff that they want you to sell to the bank customer when they come in, or one of our managers was at Harrods in London and they went downstairs to pay their Value Added Tax - they had a PC there for every clerk down there and there was one application just to do Value Added Tax - that's a perfect environment for a Java client. You don't need a full PC to go do that kind of environment and in all of these environments, the last thing you need is Explorer or Navigator. It's just the last thing you need. You just need to be able to run your applications. At home, you might want Navigator and the entire hairball and all of the features, because, at home, most people don't want their computer for productivity - they want it for activity - and that's a very, very different environment. Now what happens in the home is your business, if you're bored - buy Navigator on NT, you know a printer and the whole deal, do the whole six month deal, all right and Š that's a different environment. We're really addressing people who are trying to be productive and get stuff done and the Hot Java browser is more than capable to surf any web sites and run any Java applications, and we're going to try and keep Š our goal is to try and keep Java and the Java environment as small as possible so that it runs in an eight megabyte machine as opposed to, you know, these full-feature general purpose environments.
Andrew Birmingham:
Andrew Birmingham, Computer World. Just, I want to ask, what is it about the internet and the nature of the internet, which will preclude Microsoft from dominating it - given the resources it has?
Scott McNealy:
Sun Microsystems. I had to throw one ad in there didn't I? I don't believe that they have control over it. I don't believe they set the HTML standard. I don't believe they set the Java standard. I don't believe any more that they can own and control the interfaces. By definition, networking requires interoperability. IBM has grabbed Java like crazy. So has Novel. So has Apple. So has every one of their competitors. Everybody on the planet has decided Java is the hammer to stop the Microsoft domination - and by the way, we did it. We only launched Java a year and a half ago. Now, they didn't all sign up with us because they thought we would own it, control it, manipulate it and steal it. We have a fourteen and a half year history of doing really dumb things. We publish our interfaces into the public domain. Fujitsu is building spark chips and spark computers and not paying a penny of royalties to us and we have not sued them. Silly us. People have implemented NFS all around the world and we have not sued any of them even though they don't pay us any money. We don't believe in making money owning the language, we make Š we believe in making money in the language. Nobody owns English. Nobody owns grammar, syntax. Nobody owns the alphabet. And the net has finally won out. I believe planned Š they don't call it Red-mond for no reason - right - its a planned economy - and in theory that all sounds right, but, in the long run the market economy wins out. In the long run, capitalism wins out and in the long run the net is going to win out and open interfaces, multiple implementations, wins out. I mean can you imagine if Ford tried to move the brake pedal? I don't think so. And Microsoft is going to try and move the brake pedal. They won't be able to do it, because the users of the net far outnumber the users of Microsoft - and it is done. The rocket left the pad. The barn is empty. The gate was left open. They didn't get it. They got it. And that's just what it's all about - it's cool. Now its going to take a long time. I argue that it takes as long to decompose the hairball, as it took to compose it - and we have been dealing with them building up the desk top hairball with Microsoft for twenty years now and it will take us that long to tear ourselves away from feeling computer literate because we got it to print. We are just going to have to get over the fact that that does not make you a computer scientist. But we will get there, and the world is figuring it out. Let me tell you, its not starting at the home - its starting at business. Where CEOs are saying "Why am I giving everybody a mountain bike? Why am I giving them a 4,000 feature word processor? Why am I giving them a CD and floppy? Why do I have to use a 5 million transistor Intel chip that often does long division properly?" Š Have I gotten everybody yet? Better move on.
Valerie Had:
Valerie Had from the Financial Review. Sun took its own advice several years ago when it switched off its last mainframe and if memory serves, you had a bit of a rough period with your financials. You've given some advice today to companies not to ignore the Year 2000 problem and to stop buying PCs. How will those companies make sure that they don't go through that same financial hiccup that Sun went through when it actually went through some counter-cyclical planning in your architecture?
Scott McNealy:
Right. We actually Š Sun's kind of grown from zero to over seven billion dollars here in our fourteenth year and in 1983, or 4, we had a whole bunch of mini-computers and we were running out of space so I went to my staff and asked for a proposal on how to move forward - they came back with two proposals. One was to go on Sun and Unix to run our business and the other one was to buy a mainframe. They voted 11 to 1. I was the one who said "Let's go Sun." We voted 11 to 1 Š I am not going to ride over the hill all by myself, so we went with the mainframe. And in 1989, we were about to turn it on and they said "Oop. We need a second mainframe." I said "We haven't even turned the first one on" and they said "Well we need one to practise on." A four million dollar 'practice' machine. What's wrong with that picture? So, anyhow in Q4 of 1989 we actually turned on our first mainframe and that's when we ran into the problems that you mentioned. We had record bookings, record back Š its the only quarter in the history of the company we ever lost money. I don't like mainframes. It was really embarrassing. The Wall Street Journal was saying we needed a John Scully, kind of like Steve Jobs had at Apple. I wasn't very excited about that particular career prediction by the Wall Street Journal . We were about six or seven years building up onto the mainframe. In fact, two months after we got the mainframe up and running, they came in and said "we need to buy two more mainframes." I said "you've got to be kidding" and they basically whispered to me "you know they're only about 50 mips, right, you know, and these are not powerful kinds of environments." And so I said "Read my lips" - that was back in the days when we did that in America - "Read my lips - no more mainframes", and we have not Š I'll crash the company before we buy another mainframe and in fact I said "you can keep the two you've got, but no more" and, in fact, hopefully, next summer, we will be able to pull the plug on our last mainframe and we skyrocketed - in fact the less dependent on the mainframe, the more successful we have become - and we used to have 7% of our sales in IT costs. It is now down to about 3.2% of our revenue is tied up in IT costs. So we think it has been a very, very valuable disinvestment in the mainframe environment. How do people avoid the hiccup? I don't Š I think you can now, because I think the web is enterprise-ready. Companies around the world are using intranets. Instead of creating a Word document and sending it to 17,000 employees, we publish an HTML, a press release, and anybody wants to go see it, can go visit that site and see it. That is the way to do it. It saves your network. It saves time and you can go through the firewall and get to everyone of your customers without having to reformat it, without having to use hard copy - this is the way the world is going. This is the way to do aplets out over the network to your customers, to your partners, without having to wonder whether they're using a Mac, or a PC, or a Unix machine, or a Nortel screen-based Java-phone. This is the new world and I think it argues that you don't need any more mainframes. I'll guarantee you there aren't many companies out there who wouldn't be more successful if they just said "Read my lips. No more new mainframes. You can keep the ones you got, but no more new ones." I think it would be stunning. Its much easier to keep building a bigger hairball accelerator than it is to go re-engineer your hairball, but, I have this saying "you are all hopelessly behind dealing with the internet." You know that sounds kind of scary - I tell that to my own employees, but every one of your organisations is hopelessly behind taking advantage of the network environments and everybody is always worried about "is it ready yet?" and I argue Š there's a famous country and western song, Sometimes you're the windshield, sometimes you're the bug, right. Its Š I would always rather be the windshield and the problem was that the windshield was going too fast, the bug was going too slow. Right. And the fact that you are hopelessly behind doesn't matter. Its not what's in your rear view mirror, its what's out front. You know, I share the story about the hikers in the woods. You know, they are all hiking and all of a sudden a sixty foot grizzly bear shows up, roars up and starts running down the path at all the hikers. The Sun hiker always stops, gets their running shoes out, puts them on and as we are putting on our shoes and tying them up, people are running by saying "you can't outrun that bear in those running shoes". You say, "Exactly, I have just got to outrun you." So the whole message is here, is you have got to stop investing in the legacy PC and host-based computing environments and start building an IP network. You build the network and people will go like crazy and use it. Its like when we build a freeway in California - its clogged the day they open it. You don't stop building freeways because they all get congested. That's bad logic. People are using it and becoming much more effective. They will re-engineer your business if you build that highway.
Ian Doddery:
Ian Doddery. One of the quiet revolutions over the past decade or more, has been the growth of Unix in the commercial marketplace and throughout that time, Sun has lead that growth, and throughout that time, other vendors have claimed to be number one, and again and again, we see those other vendors setting up new organisations to fragment the drive of Unix, to create a new standard, and so on and so forth. I wonder if you could share your view of the next five years or so, of the politics of Unix and where you see the operating system going?
Scott McNealy:
That was actually a nice way of Š I get asked by the press very often Š I am going to rephrase your question in a little more antagonistic manner, you know, because people are always talking about the many different strains of Unix, and the fact that it keeps fragmenting and then we have got to go through this political process of merging and converging and it creates all this noise and energy and so, the question is usually phrased by either the customer or the press - "when are we going to get to one Unix?" I always turn that right back to the press and say, "when are we going to get to one newspaper?" And I hope its not yours - no I don't say that. The whole concept of having one Unix, is what Unix is all not about. We don't want one Unix. We want choice. There was one newspaper, it was Pravda. That doesn't work. You don't get the news that way. And you don't get innovation, you don't get the next generation, you don't get price and performance. I mean, try negotiating with Microsoft, for a better price on NT. You've taught your entire company to speak NT. Now go to Microsoft and say, "can I have a, you know, a bigger discount?" Tee hee. That's a long conversation. Now go to Sun, if we're overcharging for Solaris and you've got five, six, seven, ten different Unix flavours that are very, very identical to bang over the sales rep.'s head. That's a better situation. I tell people who want to adopt NT. I say "fire your purchasing agents." You don't need a purchasing agent, because when you decide to go with Word, you've decided to go with Windows, you've decided to go with Microsoft, you've decided to go with Intel. Do you know what your purchasing agent has a choice of? Which power supply to buy. Do I buy a Compaq power supply or maybe I'll Š I kind of like the new IBM power supply, I think I'll go with that one. Right. That's not what a purchasing agent is there to go do. The choice is what matters, and the beauty of Unix is that it has been innovating and its people have been adding new features, and opening up the interfaces so that other people can implement and do better on that, and that's why Unix has been predicted to be going away and dying every year, and has been growing like a rock every year. I mean, I hope Unix is the same disaster over the next fifteen years that it has been for us the first fifteen years. And it is all about this whole thing Š the market economy in theory is a disaster, but in practice, is the only right answer. The planned economy, where there is one answer, in theory, is perfect - one shirt maker and one button maker Š right? But in practice has failed every time. We cannot do the existence proof on capitalism, but it just works, and that's what Š that's the whole idea behind open interfaces and Unix has been kind of the kingpin, the lead warhead, and in fact, has been a totally misunderstood technology from that perspective by the customer and often, by you all, in trying to write the story - because it's a subtle concept. Its important - you have got to try and draw an analogy to but, the beauty of it is, its like a rope that keeps growing twines and then we do a political process to braid that back together into a very strong and innovative thread of technology that has migrated and moved forward very, very nicely and very aggressively - so I hope we never get to one Unix, although I mean, it would be wonderful if Solaris was the only one.
David Penna:
My name is David Penna from Amdahl. I sell large mainframes.
Scott McNealy:
And you also sell lots of Sun computers too.
David Penna:
I went to Darth Vader's presentation, sort of, down in the mouth, thinking that the mainframe was dead and he convinced me otherwise - I thought that with all the repositories, there was yet a market for us. I walked in today and been shot in the foot again. My question is relating to Unix, and when you see Solaris and Unix in general, getting large enough market share to be able to combat the growth of NT. You know, one is growing on a larger base than you are. Is there sort of a pattern that you see for the future and that sort of stuff?
Scott McNealy:
There is really Š the operating system's market is broken up into several categories. There is the server category, and there I would not want Š I would not want to run my server room on NT. I want to run it on Unix. I want something that scales, that is reliable, available, serviceable, that has network management tools, that doesn't have that feature that all Microsoft products tend to have - which is the random, daily, crash. I mean you're all using Windows on your desktop. Doesn't it do that? About once a day? You don't use it much then. This other guy has a different opinion. You cannot run Š in fact, this is where the mainframe companies like Amdahl and IBM are actually going to see an opportunity, because the server room becomes very important in this Java Client mode. And this is where you need mission critical, available Š like when you pick up the telephone and you get dial tone, if you don't have dial tone by the time it gets to your ear, you're angry, you know, you're angry - if you turn on your computer and it actually works, you get on your knees to the Pacific northwest and say, "thank you Bill". So what we are now trying to do Š a telephone without a switch is useless. A set top box without a head end is useless and a Java client without a server room that is up and running is useless, and this is where Amdahl and IBM are in fact, becoming very important allies and partners with Sun Microsystems in that world, and I just wouldn't want to be running my switch room on NT. You have got to think about that, and super bowl ads are not the way to get the MIS directors comfortable. On the desktop, there is the power desk outfit, the creators, the CAB developers, the software developers, the trading floor and again, that's where the power of Unix has won out in and is still winning out and is the dominant environment Š where we haven't is on the mere mortal - I call it the passenger car market if you will, Microsoft NT and Windows and Win 95 are the dominant environments. We are giving you a better way. Not a better OS - no OS. A Java virtual machine. The answer is not to try and do a better NT, or to try and move Solaris into the volume desktop environment, the answer is to eliminate that - its to take it out of play and actually use the Amdahl server environment more aggressively, so I think, I think the two companies are actually very well and highly aligned, in fact the entire Fujitsu group is very aligned with this, and in fact, IBM has, all of a sudden, become one of Sun's most important strategic partners, because of our view on this network computing model. So anyhow, I thank you all for putting up with my ranting and raving. Thank you very much.
LAURIE WILSON:
Scott McNealy, thank you very much for appearing here today. I must say, as someone who owns a three-year old PC and probably uses no more functions than I can count on the fingers of one hand, I used to feel something like a luddite - now I know you've restored my self-esteem. I'd like to present you with this little gift to mark your appearance here, and its something that I suspect you might actually appreciate - it's one of those old-fashioned communication tools, called a pen. Thank you.
See also:
Provided on the Web as a community service by the Australian Computer Society. Design by Tom Worthington.