PROBLEMS WITH JAVA DEVELOPMENT : : ZD Internet Magazine (ZDIM): Developing mission-critical applications on Java remains a daunting task, do you expect some of these performance problems with Java to be resolved soon?
Schmidt: That's exactly as we expected. There are many limitations of Java as it stands today. It's not fast enough. There's a whole bunch of libraries that are missing. So the primary use of Java today is for client front-end applications. You have an existing application in a client-server architecture where there's a proprietary client and a proprietary server. What's happening is that the client is being rewritten in Java, the server is not because the server software is not robust enough yet.
So the first use of Java is to make a bigger market for existing applications. If you use Java, it will run everywhere. It's easier to manage. Java is a huge success if you're doing network and systems management where you use Java to remotely manage devices. It will take a much longer time to see what you traditionally describe as server-based applications. It takes time.
In this calendar year and into next calendar year, the application that you see will be front-ends only. And it will be the following year when we begin to see the server applications. We have to be realistic about this. The other thing is that Microsoft will be shipping 50 million copies of Java next year, which will become part of Windows 97. The vast majority of the upgrade will be in the client, not the server. We expect the majority of the apps to be client-focused, which speaks to the volume. There's no reason that Java shouldn't do well on the server. But it just takes longer.

CRITICISM OF ACTIVE X : : ZDIM: Speaking of Microsoft, do you think that their adoption of Java was based on fear, not on necessity?
Schmidt: They told us that they were impressed with our success in establishing Java as the next big thing on the Web. Because of that, they had no choice but to license it, instead of building a clone. Cloning Java is difficult. And we were happy to do a license with them and I've been happy with them since then. They have fulfilled their terms of the contract with virtual machines that are very good. They're busy evangelizing for their developers in the use of Java. So far, it has been a good relationship.
I do want to say something very clearly that Active X is really a trap. I was not aware of this until I studied this more carefully. The way Active X works is that it is a library of interfaces. The library of interfaces and implementations are [built] around the network. But when you invoke and manage the Active X components, you need something called an Active X container. There's one container available. And wouldn't you know it, that container is Internet Explorer 3.0.
If you are using Active X, you're essentially forcing your customer to choose a browser. Maybe your customer doesn't want his choice taken away. I would be very clear on this that it's a slippery slope. As long as you stay on Java and the Microsoft platform, you're fine. The moment you're plugging into Active X, you just lost all the benefits of Java and you're locking your customers into a position you don't want them to be in.
You may be giving them some serious headaches. The odds of this being truly, truly open is pretty low. Active X is highly, highly integrated with Win32 and Com and Dcom, which are the underlying interface libraries. I would be extremely surprised if it's easily implemented on anything other than Windows. Again it looks like it's a good marketing tactic. The de facto action is to have a proprietary solution.

ZDIM: But the argument from Brad Silverberg, who is in charge of Microsoft's Internet platform and tools division, was that Netscape's IIOP was just a vaporware that seeks to confuse the marketplace, whereas Active X has already become a $250 million business with lots of room to grow for developers.
Schmidt: That's the equivalent of saying you ought to use C++ because all these people are using it on Windows. There's a principle involved with cross-platform flexibility. And Brad is busy trying to sell Active X, which is only available on Microsoft platform. It's a lock-in. It makes sense. If you're a monopolist, you own 80 to 90 percent of the market, this is what happens. I don't think people want to have their choices taken away from them.

ZDIM: Do you think IE 4.0 will be even more apparent in supporting Active X?
Schmidt: Of course, Microsoft's strategy - and they have been very clear on this - is Active X. Their entire Internet strategy is based on Active X. If you don't want to give your customers choices, you give them Active X (which only supports Internet Explorer). It's simple. At least with Java, Java is not as complete as we'd like it to be, (but it is multi-platform.) We're working as fast as we can. Certainly there are a lot of companies that are helping us do that. That's the pitch. Seems to work. People are very receptive to Java.

ON SUN'S RELATIONSHIP WITH IBM : : ZDIM: You got Java OS, Javabeans and so on, so how many other ways are there for Sun to skin this Java cat?
Schmidt: I don't think anyone else is complaining that we're not trying things. We have the Java platform. We have the Java workshop. We got Java OS. We got Java chips. The only thing we haven't done is Java applications. We've got third parties doing that. I think we're maxed out. To me, the most impressive thing is the partnerships. IBM is spending more money on this than anyone else I know. Who would have thought that? IBM gets this Java thing really well.

ZDIM: Do you think they could salvage OS/2 with Java?
Schmidt: I hope so. They are working on Java for OS/2. They do tremendous enterprise library work. They are committed. Look at Netscape. Netscape is doing more and more Java-related work. And we just announced our plan with Apple.

ON JAVA REVENUES : : ZDIM: What about the revenues coming from Java?
Schmidt: The good news is revenue is growing very rapidly and the bad news is expenses are also growing very rapidly. We're not going to make money from Java directly for many years. We're running Java at a loss.

ZDIM: Not even by the year 2000?
Schmidt: I can't predict that long. But I think it's unlikely. Java is so important strategically and it's worth it for Sun to invest as much as we can. It's such a paradigm shift in the industry. In terms of our servers and Solaris, we already know our sales are stronger because of Java. We call that leverage. There's no question that there's leverage. If you look at Java revenues solely, it's years away. We joke that we'll be making most of the money with a line of apparel and trade shows.

ZDIM: How much have you invested in Java?
Schmidt: The number of people working in the company on Java is many hundreds. You can conclude that our investment in Java is in the many tens of millions.

ON WHY JAVA IS IMPORTANT : : ZDIM: Some developers suggested that Java remains at best a programming language. Most users don't care whether it's written in Java or C++. Is that a fair statement?
Java is a revolution to the people who use it because they will create applications more quickly. It's going to make the market more competitive. Java as a language, we believe, will deliver twice the productivity for programmers. The reason Java ultimately is going to take off is that programmers feel that they're more productive. It's like we're giving artists better paintbrushes. They make better art, it's better art and faster. That's our contribution. It's not a huge contribution. In the whole scheme of things, the artist is more important than the tools. What really matters is whether the key software designers will be able to use Java to build innovative apps and build them more quickly than they otherwise would have.

ZDIM: And you think the performance of these applications will vastly improve once more compilers are available?
Schmidt: There's no architectural reasons that Java should be slower than C or C++. It's just a matter of time. It takes time for some of these tools to port over. Maybe it's a few months. You're competing with 15 years of mature C and C++ technologies. It's happening very quickly.

ZDIM: Is cross-platform compatibility a minor stumbling block?
Schmidt: Java solves that problem. Our applications writers are selling into a network that is mixed. If you believe your customers are only using Windows, you're going to use Active X. But among enterprise customers, the networks are always mixed. I have yet to find an enterprise customer that has a homogenous network. They always have some Macs, PCs, mainframes and OS/2 machines lying around. Java is the only answer. Part of the reason IBM likes Java so much is that Java is a way to unify all platforms. They are selling AS/400 servers and OS/2 servers.

ON INFIGHTING BETWEEN SUNSOFT AND JAVASOFT : : ZDIM: There are anecdotal reports that there's some infighting between SunSoft and JavaSoft, can you comment on them?
Schmidt: At any big company you are going to find some internal discord. The precise reason that JavaSoft is separate was that it would focus on Java, not Solaris. If we had put JavaSoft inside SunSoft, you would have gotten Solaris out of it. The important point is there's a focus at Sun for Java and that is JavaSoft. SunSoft is an ally in that. There are development tools from them. Remember that JavaSoft works with partners like Borland, which competes with SunSoft on Java Workshop. The structure has to make sense. I think it's the right answer. True, you're going to have some internal tensions. But Scott McNealy (Sun's chief executive officer) has made it clear what the priorities are. We're pretty focused on Java these days. There are not that many people who are not pushing Java at Sun.

ZDIM: So you don't think some of the people working at Solaris would feel slighted?
Schmidt: I hope not because the people at Solaris are trying to make Java run really well on Solaris. One of our objectives is that Solaris will be the best platform to run Java.

ON JAVA FUND : : ZDIM: What can we expect from Java Fund? Was this a knee-jerk reaction to perhaps not enough funding going around for Java-focused developments and start ups?
Schmidt: I was involved in setting this up. I can tell you that it was not a knee-jerk reaction. We began setting up the fund with a meeting with Kleiner Perkins in October 1995. So it had been in discussion for more than 10 months. During that time, the question was how do you build a fund that promises good returns for investors. There was no question that it was a good idea. So in 10 months, we raised $100 million and the money was sitting in the bank. They got companies like Sun and Cisco to invest. It's one of those things that covers a much longer period. You wouldn't know the fund would be successful for five years.
When you have multiple investments like this, you get a community of people and they can learn from each other. It's good that you have a community of people who are co-dependent around a core technology. Fifteen years ago, if you had a fund on Lotus 1-2-3, you could have drawn off the success of DOS. Everybody was using DOS and they could learn from each other's success in DOS. But the timing wasn't ready. It was a narrow market.

ON HOW THE INTERNET WILL CHANGE THE BUSINESS WORLD : : ZDIM: How will the Internet change the business world?
Schmidt: The rule of the network is that one plus one equals three. You and I meet each other over the Internet and we do business in ways we wouldn't have done if the Net didn't exist. For your business audience, the most important thing is that the Net brings them a new set of customers that they would not have otherwise received. That's why I think people should be excited about it.

ZDIM: Will that result in disappointment?
Schmidt: It depends on whether you have a realistic view of the Net. I have a pretty straight-forward view. The vast majority of the use of the Internet over the next five years, or perhaps my lifetime, will be communication and chat. Video will be included in that. The reason is that this is what people do. To argue that this is a brand new market is to understand that the market is first and foremost a marketplace of communication and ideas and only secondarily a marketplace of products.
If you think this is going to displace the existing market of products, then you're wrong. If you think this will become one of the primary vehicles where people select and communicate information, then you're right. I have very realistic expectations of whether people are actually going to be buying things, as opposed to finding out things. The distinction is crucial. There's no question that the Net really does change the way you market and communicate with your customers. And it doesn't necessarily mean that you're going to sell them differently.