Computing’s Second Punic War
March 1 1997 by JP Donlon
That Microsoft is rapidly extending its hegemony over the personal computing industry is not in doubt. Like ancient
Colliding with this force is the rapid acceptance of a new object-oriented computer programming language introduced in 1995 by Sun Microsystems called Java. Originally viewed as a clever bit of Web flash, it has evolved into something far sturdier and ideal for Internet and intranet computing. It has fired the imagination of both the computing industry and business users who see it as a path to lower-cost desktop computing and overall cost-effectiveness of enterprise networks. Since Java was designed to break the link between the application and the operating system, developers can create programs that run on any computer regardless of its operating system. No more separate programs for Windows, Mac, Unix, etc. While extending the life of legacy systems, it also opens the door to Network Computers (NCs), so-called dumb terminals, cheaper than PCs, linked to servers. Instead of installing new programs on every PC as is done now, programs can be installed once on a central server and are available for use across the network. Java-philes assert this will significantly reduce corporate computing costs, arguing that the lifetime service and administrative costs of NCs are much lower than PCs.
If the network really is the computer, as Sun Microsystems’s CEO Scott McNealy argues to anyone who will listen, then the battle for computing’s future course is engaged. Like Hannibal, the great Carthaginian leader, Scott McNealy has sent his Java elephants over the
“It’s great that there’s somebody willing to fight Bill Gates,” says Randy Fields, president of the Park City Group, a Utah-based information systems firm, “but I don’t think it’s winnable. Nonetheless, it’s caused Microsoft to change its strategy.” Scott McNealy, 42, is not shy about taking on Gates’s empire, nor about being controversial. The Windows operating system, he says, is nothing more than a “hairball on a desktop.” With his toothy smile and fraternity boy appearance-he almost never wears a tie-McNealy is not the techie nerd of silicon legend. He studied economics at Harvard, earned an MBA from Stanford, and counts his father, a former American Motors executive, as a mentoring influence in his life. In 1982, McNealy and two other 27-year-olds founded the Mountain View, CA-based company which has grown to be a $7 billion maker of computers and software. Direct and plain spoken, McNealy encourages Sun employees “to have fun.” Indeed, the company’s April Fools Day pranks are legendary. On one occasion, McNealy found his car mounted on blocks suspended over the water fountain in front of Sun’s
The Javameister’s challenge is to figure out how best to capitalize on Sun’s current advantage. The company derives most of its growth from high end Unix workstations and servers, but is reported to be taking a drubbing from its low-end workstation line. (Second quarter earnings surged ahead 41 percent showing improved margins, but the company cagily declines to break out performance on Java.)
Can Java and the NC humble Windows as Microsoft toppled IBM’s mainframes? It will be a neat trick persuading people to give up their precious PCs. And no doubt Microsoft will not stand by idly as Sun steals away Compaq, Gateway, and HP users. For the record, Hannibal, who as a youth swore eternal enmity to
HEADING INTO THE NEXT PHASE
Are we witnessing the beginning of the end of PC-dominated computing and the rise of Internet-based computing?
I think we’re heading into the third, most interesting phase of computing. The olden days were about host-based computers-the centralized hairball style of computing-where you put everything, all the applications and data, on the mainframe. Then you put a dumb terminal on everybody’s desk that could access the mainframe. But it wasn’t very responsive and it ran slow. So they went out and put a personal mainframe on their laps-it’s called a PC-and gave everybody their own file system, their own disk farm, their own 32-megabyte, 32-bit multitasking, multithreaded operating environment with 10 million lines of code on the Intel Pentium chip. And all of this just to type our names or to send an e-mail-classic hairball on the desktop.
That’s the PC model. Incredibly expensive, 10 to 12 or 13 thousand dollars per year. And, what step forward in user productivity have we seen? Personally, I got so frustrated with clip art and presentation graphics that I banned Power Point from our company 10 months ago. Our earnings have skyrocketed and our stock price has nearly doubled since that time. I have seen absolutely nothing but productivity gained by banning word processors with more than four features and Power Point-like graphics, or presentations graphics programs. I think if General Motors could ban Microsoft Office from their company, they could get the 1999 cars out next year at half price.
The third model is network computing. We’ve said for many, many years, that the network is the computer. This model says you use basic ASCII to create mail, because every computer in the world can read amortization, and the systems administration cost. We think by going to Javastations we can cut that in half.
Is it your aim to have Java become the definitive industry networking standard?
We want it to become the language of network computing. COBOL was the language of mainframes; FORTRAN was the language of mini-computers and technical computing. C and C++ were the right-sizing languages. Visual Basic is the language of desktop hairball PCs. And. Java is the language of network computing.
THE GATES FACTOR
Assuming that Java becomes the de facto language of computing in the future, how does Sun Microsystems directly benefit?
Today, in a proprietary Wintel world, only two companies-Intel and Microsoft-get to participate in providing the microprocessors and the platform for most desktop computers. In this new world, everybody can compete for every seat, for every client environment, and for every server environment. We’ll be able to compete. Granted, we’ll have lots of competition, but it’s better than no competition. I’d rather compete and lose than not compete at all. We’ll grow just fine in this new market. We’ve grown fine in the limited market. In 15 years, we’ve gone from zero to eight and a half billion dollars, in a tiny, little limited market. So imagine what we can do.
You’ve caught Wintel off balance by establishing this name for yourself. But you’ve also put yourself in the cross hairs of Mr. Gates.
I’ve been doing that for about 15 years. It’s fun. Everybody on the planet is in Microsoft’s cross hairs. The banks, the content providers, the network services providers, the Internet service providers, the software developers, the computer makers. Everybody out there has been in their cross hairs, so that’s nothing new. He’s got a planned economy, where he owns and controls and drives the strategy. I believe in sharing in the wealth and sharing in the innovation and making the market bigger. That’s what market economies are all about. It’s like Ford and Chevy. Coca-Cola and Pepsi Cola.
What about this so-called Net PC-not quite a PC but not quite a NC?
I call it a PC in a corset, with the wires pulled real tight and a real red face. It’s still a chubby little computer under all of those wires. And a low-cost PC is another oxymoron. So as long as they’re fantasizing that way, we’re in good shape. We’re going to have a lot of open-field running.
Allegedly, there is software that can reduce PC administration costs substantially as a near-term fix. Do you not regard that as a viable option?
Well, I would definitely try and do that, but I wouldn’t allow any upgrades. I mean, moving to Win 95 or to NT is not going to solve your problems. It’s only going to give you a bigger hairball on people’s desks. Why pour money into the wrong architecture? The model we’re proposing is one that’s very tightly integrated with the Windows environment.
Does Java in a way disintermediate software creation as it is?
It maybe disintermediates some software delivery. Like, instead of Sneaker Net-running around handing out floppies-or instead of shrinkwrapping it. Netscape is the great example of that. So much of their software is just clicked down over the network. They also shrinkwrap, but it certainly augments network-based. I don’t get any of my applications physically; I get them all electronically. I can bring 300 different applications to my desktop right there, just a mouse click away. I don’t know where they come from and I don’t care. They just show up.
A PLAN FOR THE FUTURE
Where do you go from here with Java and all its forms?
Last week we announced the 100 Percent Pure Program. What we want to do is make sure that anytime you write a Java application, it’ll run on all Microsoft computers, IBM, Apple, Unix computers, Sun computers, NCs, and Java stations. And the real key is, we’ve got to protect that. Microsoft wants to hijack it. They try and get you to use CaptiveX extensions in your Java program, so that it will only run on Windows. But if you make it work better on Windows but it only runs on Windows, then you’ve broken it. Because porting is not a useful activity. If every one of us had a different written and spoken language, we’d have to re-port everything anybody ever wrote. The fact that we’re all on one common language is wonderful; translation is a disaster. It’s what causes more wars than anything, right?
Several consultants had high praise for Java, but a few did say they felt that Java, to some degree, was still immature. What’s your response?
In terms of security, it is infinitely ahead of Windows from a maturity perspective, because Windows will never get to that level of security. So you’ve got to be very precise about what you mean by immaturity. Granted, there are some things, like Java Wallet technologies that we just announced last week, which are becoming available in the next 60 days. Yeah, they’re not there, but they’re not there in the Windows environment either. I think the maturity question is just kind of a thoughtless, off-the-cuff statement to make, just so that they sound like they’re adding some value on this thing. But they aren’t.
The [NC] has wonderfully complicated and exciting, but doable implications for every company out there. And it’s not good news for Microsoft. They make their money shipping desktop operating systems. This new model says the desktop operating system’s a negative value-add.
When we see greater deployment of the network-centric model, will software and software development be cheaper?
I think so, because today you’re paying for the cost of porting and you’re not getting the economies of scale. Look how expensive Mosaic is. It was done by a couple of kids at the
But what would be the incentive for software creators?
They all want to be Marc Andriessen, who’s a 100 millionaire right now. Maybe not a 20 billionaire, but I know a lot of people who will do a lot of things to become a 100 millionaire. I know a lot of people who’d do a lot of things just to pay off their school debts. The incentive to work hard for $1 million or $10 million or $100 million is just the same incentive as to make $20 billion. At some point, you get past working hard for that reason. You’re working hard for a reason that has nothing to do with money. Most people probably hit that at $5 million.
What are some of the exciting things that you are working on that we might see in the future?
We’re building our coverage model better, because we’ve been too small to call on everybody. We’ve now got an incredibly compelling answer, but we’ve got nobody knocking on the door. So we’re trying to just build the coverage models on a global basis, to allow us to go call on the global 1000 companies out there, because they’re the ones that really have the most to gain.
Are there industries that have migrated to the network model particularly well?
Pockets. The trading floor has done it well and is moving aggressively because the ROIs are so high. The engineering workstation world is much farther along in this network model-not necessarily zero administration, but certainly on the network model. The telephone and cable companies have driven the zero administration model particularly well, as have all the utility models, the electric companies, and the water companies.
Other than Java, what financial or other goals have you set for yourself?
I believe goals only limit you. Because if we’d set a goal 15 years ago, it probably would have been to be a 100 million dollar company after 15 years. And I would have been thrilled to death with that. And I’d have set a personal goal that I wanted to get comfortable at making $2 million dollars. So, I learned a long time ago, goals can only limit you. So my only goal is to come in and work as hard as I can every day, go home and spend a little time with my little boy and my wife, pass out, and then come back and do it again the next day.
Is that really how you manage?
That’s how I personally manage myself. We obviously have a strategic planning process where we’ve committed to our shareholders to goal our company to grow at least 15 percent per year in earnings per share, as a floor, for 100 percent payout. We certainly pay our management team to do much better than that, and bonus them accordingly. I don’t know what our 10-year EPS growth has been, but I think it’s significantly above 15 percent.
In your strategic plan, do you have any sort of long-term goals to get into other businesses?
No. The network computing business is great. We don’t need to get into components like silicon or subassemblies like disk drives. We don’t see a need to get into printers or into other computing models, like PC or mainframe host-based models. There are people doing all those things for us in a partnership role and what we really want to do is focus on providing the best client and server technologies for network computing, with open interfaces.
How do you see this evolving over the next 10 years?
Well, I’m certainly not going to rest until I know little Maverick can grow up in a world where the written, spoken language of computers is open, like English. I think the world would be a better place because of it. Every proprietary computer company has had to face the grim open reaper, except Microsoft and Intel. And I think they’re staring him right in the eye right now, trying to play a game of chicken. And so far, the grim open reaper has won every battle and forced us all into being open. Now, it just happens that we were born into that religion day one. We just said we’re going to publish all our interfaces day one. We faced the reaper the day we incorporated.
What do you make of all the Java-related commotion?
I think we’re on to something. I was at a hockey game with my little boy Friday night. Sports Channel occasionally pans the crowd and picks out people. They almost never acknowledge who they’re shooting unless it’s Sylvester Stallone or Madonna or Michael Jordan. Suddenly, the camera was right in front of us and the red light came on. And the Sports Channel announcers say, ‘There’s Scott McNealy, a well-known season ticket holder for the San Jose Sharks. And there’s his little boy-what’s his name? I think his name’s Java.” It really hit me then that this thing had gone about as mainstream as you could possibly imagine, that they would be actually yakking about that on a Sports Channel broadcast. We’ve just been plugging along doing our deal here and trying to open up computing models, and suddenly it’s become coffee-table conversation.