Digerati - Chapter 21

Digerati - Chapter 21

Scott Mcnealy [10.1.96]

Chapter 21


Scott McNealy

THE GENIUS (W. Daniel Hillis): For the last decade, I have heard nearly every year that Sun was facing a great problem. Somehow it has always managed to stay ahead. Now it's clear that Sun and Scott are flying high and everyone is wondering how they do it. They did it by hiring smart people and letting them take risks, and they moved fast enough to stay ahead of their mistakes.

Scott McNealy is the cofounder and CEO of Sun Microsystems, Inc.

"I like to compete," Scott McNealy says, "by the rules. Fair and square. Toe to toe. I don't like my competition, I'm not elected to like my competition. I'm paid to deliver a return to my shareholders by following the rules of business, by following the rules of commerce, by following the rules of the local land, by staying ethical, moral, and legal, but burying my competition. That's what I get paid to do, and that's what I love to do. A good clean legal body check is as exciting as scoring a goal."

I was predisposed to like Scott McNealy, who has been CEO of Sun Microsystems since 1984. When I entered his office and saw him surrounded by hockey sticks and other memorabilia of the sport, I realized that he was a major league hockey fanatic. Since my goal in life as a teenager was to play right-wing for the Boston Bruins, Scott is someone I can relate to. But am I intimidated by him, sitting in front of his impressive array of hockey trophies? No way. I like a challenge; I'm ready for anything he's got.

These days, Scott has a lot more than hockey on his mind. He's the driving force behind the leading network-computing company, Sun's bread-and-butter business. More surprising is his role as leader of a team of world-class computer scientists who distinguish themselves by making a significant contribution to the state of computing as a whole, as well as to Sun. You wouldn't expect a Silicon Valley body checker to carry on in the R&D tradition of Bell Labs, Xerox PARC, GE's Sarnoff Center, and IBM's Watson Center.

Scott claims, however, that he is not a visionary but rather a good businessman who stays focused. "I leave the vision thing to the gurus and scientists," he says. "What I do is articulate what our people think is the right answer." There are plenty of visionaries and gurus to choose from at Sun. Scott can look to Bill Joy or James Gosling or Eric Schmidt or John Gage or Lew Tucker‹just to name a few. His role is to interpret and articulate what the visionaries are saying, evangelize these visions, and put together the resources to take advantage of the discontinuities of the "paradigm" shift caused by a new technology which happens in the computer industry every few years. "I'm on the bully pulpit," he says. "Fundamentally, the CEO's job is to figure out what the vision is, not necessarily create it. Develop a plan that uses company resources to best take advantage of that plan. Get it approved by the board, then go on and execute on it, deliver the numbers to the shareholders, and get yourself reelected another year. That's my job. I decide who's on my staff, I charter them, and I approve the plan, and away we go. Then I spend the rest of my time evangelizing‹where we're headed and why it's the right answer."

The current prominence of Java‹a programming language for the Web‹is a good example of the benefits that Sun derives from supporting internal research efforts. Java, originally developed by Sun engineers in 1991, is a highly interactive language that allows users to download small applications‹called applets‹and run them on any type of computer, using the "Java Virtual Machine." Java is the first "network-smart" and "platform-independent" programming language around. Though the Internet has been important to Sun since the company was founded, until recently the Net was not seen as the main plank of Sun's continued success. The explosive growth of the Web offered opportunities that the company has been able to vigorously exploit. Java is becoming the standard for Web programming, opening up completely new market opportunities for Sun. By changing the way in which applications and content are written and delivered to desktop computers or "clients"‹across a network‹Java offers the possibility of altering the entire competitive landscape to Sun's advantage. In this regard, Scott is going after what he refers to as "the desktop hair ball" of computing and he clearly has Microsoft in his sights.

Scott McNealy is "The Competitor."

THE COMPETITOR (Scott McNealy): At Sun we believe in the network-computing model. We're not wired up and married to the host-based centralized computing model, and we're not all tangled up in the desktop hairball‹that is the desktop computing model of the Intel-Microsoft world. Everything from the first computer we shipped a long time ago goes out with a network interface, and every desktop, server, application, software product, and service product that we've ever offered has been network-centric. That's probably our biggest advantage. Our second biggest advantage is that we own all the implementations of the key components based on openly published interface specifications. We have our own microprocessors. We own the user interface and the networking interfaces, in the sense that we implement TCP/IP, Corba [common object request broker], and all the other network protocols. We have the technologies that allow you to run your network environment. Then we have servers and desktops to help you go out and deliver the value of the network computing environments. We have all the pieces that really matter to the customer.

There are two kinds of companies: product companies and trading companies. Sun is a product company. Our integration services and capabilities are all focused around network computing and our products. We are not like a trading company, which services any product, or writes software for any product, or trains you on any product. We do software development, we do integration, we do training and consulting on our products, around our products, for our products, and we help you integrate with other environments. We sell in highly focused markets, which are tightly related to each other. We do not do the services for other companies' products. Only IBM and DEC seem to want to do both. Everybody else is in either the integration business or the products business.

Since 1985 we've said "the network is the computer." That's been a nice little tag line. We got off on tangents and invented some other tag lines, but we came back to the original one, because it is the right one. The network is the business, the network is the future, and the network is what matters. So we say the network is the computer.

I'd love to evangelize and make credible the myth that the computer is disappearing. The more people who leave the computer industry, the better, because we're here to stay, and we're going to be one of the Big Three computer players, if you will. In the same way that the automobile industry consolidated around three players‹General Motors, Ford, and Chrysler‹the computer industry is going to consolidate. I'm not talking about the integration-trading-reseller channel. There will be lots of resellers in the same way there are a lot of car dealers. There will be only a few integrated product companies, such as Sun, that do microprocessors, operating systems, user interfaces, networking pieces, and do the final assembly and integration of desktop and server and clientside computing environments.

We want to be one of those three. Two of the survivors are what I call "General" and "Motors," Intel and Microsoft. They have all the pieces to make the Wintel computing environment. They just don't seem to work together very well. IBM will be Chrysler, going back to the government for loan guarantees on a regular basis. We will be the Ford Motor Company of the computer business, with our integration from the microprocessor all the way through service and support.

Silicon Graphics is going to make a nice division, at the very high end, of one of the Big Three computer businesses at some point. You might think of SGI as the sport-utility vehicle of the computer industry. The product line could be very profitable, but needs to be part of a larger organization. There aren't going to be niche companies in the computer business in the same way that there aren't any car companies that specialize in two-door sedans or four-door convertibles. You've got to be a broadly global player with volume. Scale really matters in our business.

Unix is the enterprise server environment for databases, for large hardware data warehouses, and for file servers, Web servers, and security servers. All of the features of Unix make it the best scalable enterprise-server environment. Unix has also won on the power desktop for CAD engineers, software developers, and Wall Street traders, because of its multiprocessing capabilities.

What has won on the corporate desktop has been the Microsoft environment. Within that particular environment, we've given everyone a mainframe on the desk or in the lap. Everything's there except the halon fire-retardant and the water-cooling system. You've got a file system. You've got a disk farm (i.e., massive amounts of storage). You've got a 32-megabyte, 32-bit multitasking, multithreaded, symmetric multiprocessing, scalable desktop operating environment, with a backup medium known as a floppy and a software distribution mechanism known as a CD-ROM player. You've got all the middleware and the bells and whistles and configurations. PCs are like thumbprints. No two PCs in the world are configured alike, and they are a nightmare to administer. If you want to keep kids off drugs, give them Windows 95 and tell them to get it up and running on your current home PC. That will keep them at home for a long period.

The challenge and the opportunity is to come up with a new paradigm for client-side computing. That's where the Java model makes sense. You can deliver a Java-based computer that has no disk drive, no floppy, no CD, no operating system, and minimal memory; then you can use the network to store files, applications, data, video, audio, Web access, security, and billing. All these activities can be handled in a server room by a trained professional. This is the way technology can be driven into a ubiquitous usage model. The obvious example is the telephone. When I give you a telephone, I don't give you a handset and a switch on your desktop and say, "Program the switch, configure the switch, load software into the switch, write software on the switch, back the switch up, and carry the switch around with you." If we were doing that, we'd never be able to make a phone call.

The new model of the Java client for the network terminal, or the network computer, is a Java Virtual Machine run in a browser environment. You just turn the machine on, click, and down comes your word processor. You create your application, send it back, and have it filed. You download video. You go out and surf the Internet. You do whatever you want to do, but when you're done, everything is stored and managed in the server room. You never run out of disk space, and you never have to worry about your battery running low. You always have a datatone.

Think again about the phone. If you pick up a telephone and don't get a datatone by the time you get the handset to your ear, you're angry. Contrast that with turning on your Wintel computer. A little part of you goes, "Yes, it booted. It's working today." When it crashes every day, which it does, and you can recover something from your file, you're thrilled to death. When was the last time you got cut off during a phone call, other than cellular, and how does that make you feel?

That's the difference. We believe that the datatone model of computing, where the hard work is done for you, is the model that allows you to execute content on your desktop. You still get the power of a microprocessor, dedicated to you. With Java you don't have to buy Intel. You can buy a chip that does long division properly and has the Visual
Instruction Set. You can buy four of them, strap them together, and really have some horsepower on your desktop.

You even don't have to carry a laptop. You check into your hotel room, and there's a network workstation sitting next to the fax machine or part of the fax machine. You dial up, call your server‹either in your company or at your service provider‹and log in to your server room. You load your applications. You don't have to worry who made that Internet terminal, because as long as it's run on the Java Virtual Machine, your applications will run, and you can browse. You can do everything you can do on your own computer. The problem today is, if you use Unix, you can't use Mac; if you use Mac, you can't use Windows; if you use Windows, you can't use the other two. But in our environment, one computer does it all.

The Java phenomenon creates a couple of challenges for Microsoft. First of all, the Java client says that for most computers users, the desktop operating system is a negative value-add. There's no reason we mere mortals should have to mess around with ten million lines of code just to type our names. There's something wrong with that picture. A big part of Microsoft's net worth is tied up in keeping people dedicated and committed to the desktop computing mode. The other challenge is that in Java, you can write a desktop productivity tool in what I call "subset" mode and then publish it on the network for free.

By contrast, the "superset" model is represented by Microsoft Word. One CD, one stocking unit, one tested unit, is available for everyone to do publishing. Every known feature on the planet you can possibly imagine is embedded into it, and it comes with a stack of manuals. You have to get in the car, go to the store, buy it, bring it home, unwrap it, stick it in, and download it into your machine. Then you've got to figure out how to work it. You've got to have a superset of every possible known feature.

Think about the Java applet role. When you're on your network client, your Java client, you click on Word Processor and download a four-function word processor. It comes down in a heartbeat, because there aren't many lines of code. It has four functions: Backspace, Delete, Cut and Paste, and Print. If I need right-hand justification, I can download that. If I need a spellchecker, I can download that. If I need a new font, I can download that. That's subset: an object-oriented, scalable, robust, enhanceable kind of environment.

In this new model, I can probably get that applet for free. Some really powerful author, a single individual, will write a wonderful word processor, put it out on the network for free, and get millions of users. Then McDonald's will pay this guy a million bucks a year to put golden arches around the border of the applet and get ten million­odd exposures every time anybody wants to use a word processor. There might even be a "click here if you want a burger" button on the word processing applet. I'll deal with that for a free word processor rather than spend hundreds of bucks a year with Microsoft. I get the kind of application that I can use and that doesn't need documentation.

This puts huge pressure on Microsoft's billion-dollar apps business. With Java Office or Java Word or Java Excel potentially out there for free, or near free, we're challenging Microsoft to buy into Java. Microsoft is huge and has much to do: launch satellites, buy the Bettmann Archives, get Marilyn Monroe locked up and proprietarized. They must be seriously trying to figure out how to support a $56 billion market cap. I'd be a little agitated, too, if I had to figure how to do that. I think we've got upside here.

Every device that has a microprocessor and a network port will be an IP address on the Internet‹they will have connectivity to all the devices and networks on the Internet. They will run the Java Virtual Machine, will be able to execute content, will be able to browse the Internet, and will be able to download content from over the network, anywhere, everywhere. This means that every hub, router switch, printer-copier, set-top box, game machine, nomadic computer, desktop computer, server, automobile‹you name it‹will have a network port. You can download a game to your kids in the back seat of the car, when they get bored with it, you can download another one. When your car doesn't start, you can download diagnostics, or checklists, on what to check to make your car start. This kind of executable content will be downloaded to you at these different IP addresses on the Internet. Today, on average, we have less than one IP address per citizen on the planet. With our model, people will each have a half-dozen or a dozen IP addresses. Every cellular phone in Europe today is an IP address. This is just an extension of the Internet.

The Web is the biggest online library on the planet, by a lot. It turns out that Web interfaces and technologies are how Sun is going to publish information out on the network, so that anybody and everybody from any machine can view it. This is a powerful new paradigm.

Inside Sun we now publish most of the information that we need to share among ourselves on our own internal Web, the Sun Intranet. We use the network to run the company. We have about 16,500 employees. We do a couple million emails a day, so that is the killer app. That's the way people communicate. I certainly like email better than I like voicemail. It's hard to speed-read voicemail. It's hard to print voicemail. It's hard to cut-and-paste voicemail. It's hard to intuitively forward and copy people on voicemail. People tend to have very little filter between brain and mouth. Because they can't type well, they tend to have great filters between brain and fingertips. So with email you get something more concise, readable, and manageable.

Every project at Sun has its own Web page. In fact, we're moving to the point where we want every employee in the company to have a homepage. I do a McNealy report radio show every two weeks at Sun, which we record digitally and put on a server. The favorite section is my pet peeve section, where I rail on about any particular pet peeve I have. Then we send everybody email. They click on it and up comes a browser page, and they click on that and listen to the audio streams. We're using the Web, Web technologies, the Internet, and email technology to run this company. I believe that the reason Sun has gone from zero to $7 billion in fourteen years is our use of the network.

THE MARKETER (Ted Leonsis): I had the good fortune at Redgate of working very closely with Sun. When I came to America Online, I saw Scott one day in the waiting room. He had come to talk about some big strategic issues, but he brought with him the local sales manager. It was great to see this CEO of a multibillion dollar company go from thinking big thoughts and strategic alignments to trying to sell us some workstations. You have to admire Scott. He's a big thinker, but he executes.

THE CONSERVATIVE (David Gelernter): On the one hand, I wonder how Sun will survive. Its market is people with a professional interest in computers, and that market is becoming peripheral to the computer business. Unix, developed around 1977, is dead. It was a great idea at the time. On the other hand, the sophistication with which Sun has exploited Java is stunning. This is one hell of a smart company.

THE PRAGMATIST (Stewart Alsop): I think of Scott as a prankster. He loves to needle people. He gets pleasure in making people angry, but not in a mean way. 

THE GADFLY (John C. Dvorak): The most quotable man in the industry. Outstanding at making interesting observations and doesn't care about the consequences as much. Spends too much time bashing Bill Gates.

THE IDEALIST (Denise Caruso): Scott McNealy comes across as the eternal fratboy, but he's wicked smart. Back in 1985, when I was covering Silicon Valley for Electronics, Sun launched a sales campaign called "The Network Is The Computer"‹which nowadays is a pretty good description of the global Internet. Whether or not he came up with the slogan, he was dead right about where the future was headed. I'll bet how right he was surprises even him.

THE SCRIBE (John Markoff): Scott's a remarkable business strategist. The Java strategy is a brilliant gamble that may eventually provide Sun with an end run around the entire Microsoft monolith. At critical junctures in the computer industry he seems to have pulled these strategies out of his hat: open systems, SPARC, and "all the wood behind one arrow" have all served to keep him one step ahead of the Hewlett Packards, Digitals and IBMs in the workstation business.

Back to Contents

Excerpted from Digerati: Encounters with the Cyber Elite by John Brockman (HardWired Books, 1996) . Copyright © 1996 by John Brockman. All rights reserved.