Kim Polese talks Java

JavaSoft's former `one-person marketing department' describes why Sun `gave away the crown jewels' of Java, identifies ripe markets for today's Java developers, and discusses Java's future.

1 2 Page 2
Page 2 of 2

Polese came to Sun from Intellicorp, a company offering expert system application frameworks. After several years as the product manager for C++ in Sun's developer products unit, she moved over to FirstPerson, Sun's "spinout" group that was working on what was then known as Oak, a computer language destined for settop boxes and interactive TV. Within a few months, Polese had co-authored a plan designed to focus instead "on where the real set of programmers and where the real information superhighway was and is: the network, the Internet, the intranet, and the desktop."

In this interview with JavaWorld, Polese recalls how she acted as a one- person marketing department for JavaSoft, helping transform Java's business model, choose a new product name, lowering barriers to acceptance, signing licensees, creating a brand identity ... in short, making the Java technology "sexy." The result: the Java brand has eclipsed all of Sun's products, and perhaps even Sun itself, which is "aligning around Java." The Java wave has not yet crested, and has attracted everyone from Netscape and Silicon Graphics to IBM and Microsoft. While Java creator James Gosling apologizes for all the hype, Polese points to all the attention as evidence of her success.

Herein, Polese also sheds light on areas ripe for Java development today (such as the MIS and financial arenas), identifies likely revenue sources for these key areas, offers a vote of confidence to JavaSoft sans Polese, and predicts Java's future.

This interview took place in February 1996, on Kim Polese's last day with JavaSoft. She's joined Arthur van Hoff, Sami Shaio, and Jonathan Payne to form a yet-to-be-named startup company focusing on Java.


JavaWorld: Tell us how you came to join what's now known as JavaSoft, and what you've done there.

Polese: I came to Sun seven years ago, and I was the product manager for C++ over in SunPro, now called DevPro [SunSoft Developer Products]. Prior to that, I worked at a company called Intellicorp, and they were doing expert system application frameworks, There I was consulting, helping people build applications. So my background is technical, computer science and biophysics at UC-Berkeley. The move to marketing and to the business side of things came when I came to Sun and became the product manager for C++.

The background in C++ was very good for moving into the job that I'm in right now. I came to FirstPerson, which was the spinout company [dealing with Java technology] two-and-a-half years ago, the end of summer, 1993, to be the product manager of Oak, now called Java. The C++ background [helped] because it's a programming language and development environment background, and [focused on] the audience that we need to appeal to with Java.

Moving from interactive TV to the Internet

I co-wrote a business plan about four months after I came to FirstPerson that said we have to stop doing what we were doing, which was the interactive TV consumer device market, and focus on where the real set of programmers and where the real information superhighway was and is: the network, the Internet, the intranet, and the desktop.

At that time it was a rather revolutionary concept, because two years ago, interactive TV was it. Everybody was building companies and products around that market. But it felt a lot like AI [artificial intelligence] to me in terms of the hype factor versus the reality factor. The artificial intelligence industry [experienced] tremendous excitement about the potential, but incredibly expensive investment for basically no apparent market or customer base. So that set off warning signals in my mind.

After culling up some numbers and really looking at the market data and projections and taking a reality check in general, it became obvious the Internet was really what was going to be launching the next wave of computing. And that was the perfect environment for Oak, because everything that made Oak perfect for the settop box environment, the broadband environment, also made it ideally suited for the Internet: the security, the platform independence, the dynamic code downloading [and so on].

Giving away the crown jewels

Sun, the executive staff, deserved a lot of credit for letting us run with this, because there was no guarantee that it would work, and what we were doing was giving away the crown jewels, and really opening up everything -- total source code, even the specification -- and putting an amazing investment of intellectual property out there on the Internet for all to download.

But the gamble was that basically lowering every barrier to acceptance in the developer market was what we needed to do. And that's what we did. Barrier to acceptance is price. OK, it's free. [Laughs] People were trying to charge royalties at that time. I really worked very hard on figuring out what the right business model was, and that time Kalieda, QuickTime VR (Apple's product) were extracting royalties for runtimes and it became pretty obvious to me that's a good way to kill a new language. People just won't pay royalties. I was very insistent about that, and also about getting the source code out there. Because promoting ubiquity means getting Java everywhere -- not just SPARC, not just the PC, but everywhere.

`Sexy technology'

There were a variety of areas that I really worked on. The business model was one. Lowering the barriers to adoption in the way that I've described, and availability, i.e., making it very easy to access by putting it out there for all to see. The other was making this technology sexy, because it really is sexy technology. But when you try to describe it based on its feature set, people's eyes glaze over. And there is so much there.

There are so many features that make this language environment uniquely designed for the network. The security, we talked about platform independence, the dynamic code downloading, multithreading for multimedia performance, the compact size of this thing so that it can fit into very small devices, and that the code, the byte codes themselves when they fly across the net, are very small, so that they're designed for a lightweight environment. The simplicity of the language compared to languages like C++. Garbage collection. We have features built in for robustness because the goal was to make a system that could run in a settop box without needing to reboot your settop box or your toaster oven, which you're certainly not going to do. A whole bunch of features, but when you try to reel them off, it's very difficult to understand what this thing really does.

Endorsement from `the god of the Internet'

So anyway, the things that I drew up were business decisions, the licensing, the pricing, the distribution, the fact that we did this on a desktop and the Internet versus interactive TV, and then the whole branding/positioning set of things. And of course business development. Getting early partners like Netscape signed up was really critical, because we were trying to do something huge here. This was not a small play. This was -- we gamble, and either we win big or we totally lose. There's no middle ground if you try to establish a standard, which is what we were trying to do.

JavaWorld: What was the evolution of the arrangement with Netscape? How did that get initiated, and how did that eventually get worked out?

Polese: I felt very strongly that they should have an early peek at what we were doing, because it was certainly very relevant to their business. And being as they were a highly visible company on the Internet, leading the way, having their endorsement was also very important. So I gave Marc Andresseen a pointer to the release two weeks before we let it out for the world. He got an early peek at it, and he loved it. And what happened next was a very nice turn of events. I also spoke to a reporter for the San Jose Mercury News, and he wrote an article. He was so impressed with Java that he wrote an article that was published on the front page of the main section -- the headline of the front page of the main section on March 23, 1995, a day I will never forget.

JavaWorld: That's probably something you never dreamed would happen.

Polese: No. We have posters made of that headline. It's one thing to be the headline in the business section. It's another to be a headline on the main section of the newspaper itself.

So the phones started ringing off the hook after that. And the thing that was really helpful there was that Netscape was very favorably inclined to it, and Marc Andresseen was quoted saying something very complimentary about it. So having his endorsement was like having the endorsement from the god of the Internet. That went a long way with people.

Don't just tell; show!

But beyond that, the other thing that we did that was very important was to not only describe the capabilities of the language: "Imagine, if you will, that this language environment could do dynamic shopping experiences." Rather, we created an application, in the form of HotJava Web browser, with a bunch of examples -- Duke somersaulting and tic-tac-toe and live financial portfolio management application -- that showed, again, in a very visual, compelling way, what the capabilities of the language were. And that was something that I again learned from my previous experiences in trying to communicate the capabilities of a technology and not being able to, because your imagination can go [only] so far. You really have to see what the thing can do. So that's what we did -- create the killer app and put it out there along with the technology itself. Show people the end result, not just the capabilities.

In May 1995, we launched at Moscone [at SunWorld '95], and it was tremendously exciting for all of on the team because the vision was finally reality. We celebrated, as you can imagine. It was just a real thrill to have weathered the storms that we had to go through and that you really have to go through when you're trying to do something big, like we were. There are always going to be twists and turns in the road.

Eclipsing Sun: the Java brand

With branding, people have a tremendous interest in making sure that their customers know that they're running a Java application. So we also have a brand logo called Java-powered that anybody can take off of our Web page and stick on their Web page or application. It's actually proved to be a very successful branding campaign because Java is the most successful brand that Sun has ever had. And I'm amazed to come across the average person, somebody who works in a totally different area -- retail clothing, whatever -- that has heard of Java. They don't quite know what it is, but they know it's really cool. So Java actually has seeped into the consciousness of some percentage of the average consumer, and that is increasing, and I think it's a tremendous opportunity for Sun as a company to move into new markets.

JavaWorld: Java may not be a well-understood concept or technology in the consumer market, but it seems to be throughout the consumer market, and Sun hasn't had the whole computer market's attention until now.

Polese: That's right. And think of "Intel Inside." The average person doesn't know what the SPECmark of a Pentium chip is. So they don't really know what that means. What they do know, though, is that it's a brand that has quality and it means that it's probably fast -- high performance. You know, some basic, gut-level reaction to the brand, that's what you want. And that's what people have when they hear Java. They think "Cool! That's really hot technology. You can do cool stuff with Java." So it's kind of almost a status symbol to know what Java is or to be able to talk about it or develop in it, which is really the ultimate goal for a branding program is to have that kind of a very positive feeling, instant reaction to it. So I've been working real hard on that.

JavaWorld: I guess it's the ultimate compliment, and likely creates some frustration in other units at Sun, that all the attention is being paid to Java instead of what they argue is still the core business at Sun.

Polese: Actually, I would say that what I've seen within Sun is that the company is really aligning around Java in their particular business units. For instance, SunSoft is building a development environment for Java, and the hardware group is coming out with Java chips. All of that kind of activity is actually reflecting the fact that the company as a whole is very excited about Java, and everybody is really getting behind it. So I haven't really seen the frustration. I've seen the excitement.

In addition to business development, branding, licensing, and getting Java in real-world applications, [I've been] focusing on the next features in the language, like where do we really want to take Java as a language. And just communicating to people what Java is and how it's different than C++ or Visual Basic or Shockwave. Or applets versus OCXes. That kind of information can be confusing unless it's really clearly explained or articulated.

`Reality is key'

Since the end-of-May release, I've been working on business development and expanding the marketing branding campaign and getting customers, getting people out there, actually understanding how they're using Java and helping them apply it to specific application areas like the financial market, manufacturing, MIS, so we can really show people what Java is capable of with real applications.

Actually, there are about to be "customer success stories" published within the next couple of weeks, profiling companies like Starwave, National Semiconductor, Market Vision, banks, Eastman Kodak, Andersen Consulting, and showing, again, in a very real- life way, what the capabilities of the technology are. That's always been a real focus of mine, again because of coming from the AI days and learning the lessons of that. Reality is key. Demonstrating, showing in a real-life way what the capabilities of a technology are. That's the way to market it. Not describing a feature set.

The first Java applications: MIS, financial, entertainment

JavaWorld: A lot of development teams or organizations as well as businesses starting to look at Java are curious about some of the examples that will obviously be presented with these stories as they come out. Perhaps from your perspective you could tell us what you see as the top five or ten types of applications where Java will find its first valuable uses.

Polese: I think the first is going to be the MIS environment, simply because it's such a large market and there are so many applications of Java inside a corporation, and because the networks exist and the security measures are less stringent, because you're talking about typically internal networks. For instance, Sun has a whole set of applications that we run to help us [view] our 401K and check our balance, and submit employee expense reports, and figure out what our health plan should be.

All of those applications would benefit tremendously from Java, primarily from a cost-savings standpoint if you're an MIS manager, because suddenly you don't have to port the application any more, and you never have to maintain old versions of it because the version that everybody is running is the up-to-date version, because there's one copy of it, there's one version, and it's running on a server, and everyone is accessing that version of the application. Huge cost savings in no more porting and no more maintenance of back releases. And from the standpoint of distribution you don't have to physically ship out a disk any more. You don't have to deal with installation problems; everybody is just accessing the application off of a Web page.

So the combination of Java and the Web for internal operations for companies is just a huge, huge breakthrough, and the MIS community is extremely excited about this and they're jumping on it. So I would say that's a huge area.

Another one is going to be financial markets, because of the Wall Street firms -- the competitive, high stakes in that environment. They have to get out there in front and use the latest technology to deliver services to their clients, and the ability of Java to provide real-time data -- and because Java is a very powerful general-purpose environment because it can do serious computing on the client side, it can stand up to the demands of these real-time financial systems that Wall Street builds. So they've been all over it.

Another obvious area is the entertainment online services. Advertising. Companies like Starwave, Time Warner, and others who need to provide very compelling, visual, differentiated services on the Net, because the way they attract customers is eyeballs, and eyeballs are there because this page is cooler than any other page.

Java is not just an animation language. It's general purpose. Anything you can imagine doing in C or C++ you can do in Java. Which means that you can do things like real-time decision support. And you can do things like interactive 3D and shared whiteboarding and extensible chat environments where you can design a room and move new furniture into the room and you're not locked into just one set of objects like you are with a typical groupware environment such as Worlds Chat, where you're locked into a specific set of objects that the engineers who designed that system decided should be there. The chair in the room, the table, and the lamp, and that's it. With Java, because it's an extensible environment that can basically download new bits of code that it's never seen before, as long they're in Java, the environment can be extended on the fly, and you can create new objects. So you can imagine a couple of architects, for instance, collaborating on designing a house, and one of them has a cool idea that he wants to demonstrate halfway across the world, and he can do that with Java because the environment is extensible.

So those three areas I would say are probably going to be the first to really take off.

Where's the revenue?

JavaWorld: When you look at those areas, where do you see the revenue? Obviously in the first case (MIS) it's savings.

Polese: Yes.

The second area, financial, is going to be in services. That's going to be where the money really comes from Advertising and services. [In the third area, entertainment online, revenue will come from advertising.]

[Java and the Internet let financial services companies] provide new capabilities to their clients. A Wall Street firm that wants to give you the ability to manage your portfolio at your desktop and get live data, see what the value of your portfolio is at a given second based on live data that's streaming in and complex calculations that are being done at the client side and see performance. That's why you would sign up for a particular firm's service, and you're paying them a monthly fee, or perhaps you're paying them per access. Those repeated fees, I think, are where much of the revenue stream is going to come from on the Net.

Best opportunities

JavaWorld: In terms of opportunities for commercial application development with Java, are there places that are already oversaturated that companies might want to avoid, or are there especially worthwhile areas that are generally being overlooked?

Polese: I would say MIS is an area that's going to explode that's sort of been overlooked as an optimal application of Java. Helping a company run its internal operations, basically. Java for the enterprise, the enterprise being businesses and the kinds of applications that they have to run to make their business fly. So it's Java for the enterprise -- that's where I think Java's really going to be talking off and where people really haven't been aware of its capabilities.

Long-term predictions

JavaWorld: What about the applications of Java in the longer term? How do you see that evolving, say, in the next two to three years?

Polese: I see Java moving into the devices that FirstPerson actually was envisioning it being in. So these are the PDAs and of course the network terminals we keep hearing about. I think that's entirely real. The potential there is not just a pipe dream, because the technology is up to it. This is not something that we have to go and design a system from the software standpoint to facilitate its capabilities. The capabilities are there today to run bare metal, run the Java system and to run in a very small address space, memory space. So new devices.

What else? I would say just from a standpoint of programming language popularity, I think Java is going to replace C++ and is going to become the language that university students learn when they sit down and write their first line of code, because it is the best example of what a programming language should be. It is an exemplary programming language. It has all the right features in it, and it's very elegantly constructed and architected.

Competition? What competition?

JavaWorld: Do you see other languages competing for that space?

Polese: No, actually I don't. And that's one of the reasons Java has become so popular so fast. There was nothing else that was really an alternative. There is no other language that is totally platform independent, that has the security provisions, that has the dynamic code linking, the transparent code-linking capabilities, that is compact, that is simple to use, that has a familiar C++-like syntax, that's multithreaded, that has robustness features that help you write programs that don't have bugs. There's no other language that has that set of features. There are languages that have subsets of those features, but then they have drawbacks, like they're too slow, or they're too large, or they're hard to use, or they are a completely unfamiliar syntax.

So basically James Gosling and the core group of original Java designers, when they were architecting this language, five years ago, they did it right. They got everything right. And there is no language that can really compete with Java. It would be very difficult to build one that does, because of just the vast number of features that were implemented in the correct way, and because it's really difficult to build such a thing from the ground up. This was hard computer science problems here that were being solved by these guys. Like the security model. That's tough.

Java improvements

There are a couple of features that we have stated that we are going to improve in the next release, and one of them is performance. Java is right now between 10 and 20 times slower than C++, and that's an issue if you are doing something that's very compute- intensive on the server side -- lots of simultaneous database accesses, for instance, or if you have something like a JPEG decompressor, some very performance-intensive graphics applications running. In that case, you probably would use C or C++. In the next release of Java, though, we're going to have these just-in-time compilers that take the Java code and compile it directly to machine code, and in fact we're going to get better than C or C++ performance because that code will be optimized for the particular chip that's running on that device. Right now, if you're running an application on a typical PC, it doesn't care whether you're running a Pentium chip or an X86 because it's been optimized for the lowest common denominator in that box. But when we write the code generators we're going to be optimizing for a specific chip set, so Java's really going to scream. It's going to be very fast.

That's one issue. The other issue is the security model. We're expanding that so that you can do things like save to disk and access files off of a disk, because right now it's quite constrained.

JavaWorld: So that will be in the next release, those issues will become moot?

Polese: Yeah. Both of those features are targeted for [late summer or fall].

JavaWorld: What is the solution that will let, for example, someone running a spreadsheet application save their file to their local disk or read a file from their local disk?

Polese: That will be an extension or enhancement of the present security model. So basically they're working right now on figuring out how to maintain the level of security that we have in the product, which is the highest certainly by a long shot available in any programming language, yet provide these features that you want when you're building a real production system.

JavaWorld: And what about the concerns for the international group as far as the munitions laws and encryption regulations prohibit some exports.

Polese: Unfortunately, that's not something that we can control, and we're hoping that the U.S. government does the right thing, like everybody else in the country.

JavaWorld: But essentially you don't have a solution to that. You have to rely on the government to open it up.

Polese: We can't get around export restrictions. That's a pretty hard wall.

Making Java easy to use for non-programmers

JavaWorld: So some of the long-term uses for Java are new devices and PDAs and also becoming a replacement for C and C++, not only in a classroom but also in the real-world business development. Are there any other uses that you see coming along?

Polese: Making Java easy to use for the average person -- that's going to be a tremendous breakthrough. Because right now, Java's for programmers. But the kind of tools and environments that can be built around Java are pretty awesome. And I think the real breakthrough will come when a non-programmer can sit down and create a very exciting online world without having to write a single line of code in Java.

JavaWorld: How tightly related and co-reliant will things like VRML or other technologies be insofar as enhancing or extending Java, versus Java alone delivering such things as 3D worlds?

Polese: I think everyone has to agree on a standard way of doing 3D for Java, for instance, and a standard way of doing database access for Java. And that's something that we are working on with a couple of other companies. In the case of 3D and multimedia, it's Netscape and SGI and Macromedia. In the case of database access, we're working with our database partners, the big database vendors, on getting Java to easily access SQL-based databases. So that's what needs to happen. There's not one particular technology I would point to. VRML certainly is one that's coming out and emerging as a frontrunner for 3D, so Java has to talk to VRML. Java has to talk to every standard that emerges out there, but I don't think we're going to be able to drive what those standards are.

Also, the other thing is that the vision of Java, and the one that I think is becoming real, is that people can take Java and hook it up to whatever their favorite media type or application system happens to be. So if there's some cool new 3D protocol out there on the Web, we have the technology out there. The source is out there. People can go ahead and integrate Java with the new system and put the results out there on the Net for other people to browse. So this is part of the openness thing. We're not going to be able to say, "Hey, you can only do 3D with Java this way. Forget it if you want to do anything else." That's not what we want to do. We want to encourage movement forward, vis-a-vis any technology. In other words, the right 3D standard should emerge, and Java should talk to that standard, just to use 3D as an example.

Java vs. Visual Basic

JavaWorld: What do you have to say to the fans or users of Visual Basic and what Java has to offer to them?

Polese: Visual Basic is a very appealing one, and there is a tremendous installed base of programmers or users of Visual Basic. Visual Basic has huge pluses going for it. The minus, the huge drawbacks, of Visual Basic are, for one thing, it's Windows-specific. It really is very platform-specific. Even [if] it does run on Macintosh, the fact is that the best, earliest versions come out on Windows and all the hooks into the other applications from Visual Basic are Windows applications. It's a Windows-centric language. Java is not an operating-system specific language. Visual Basic is.

Furthermore, Visual Basic is not a real programming language. No matter how you cut it, it is a scripting language and it does not have the capabilities to let you build a real production system. You run into trouble. You start writing spaghetti code if you start getting into any kind of whole, full-scale application. And then you have to drop down into C or C++, with Microsoft's model; then you're writing in Visual C++. Now, again, you're dealing with the fact that this is platform-specific. You're gonna have to port your application if you want it running anywhere else but Windows. Furthermore, C and C++ are not secure environments. You're going to be sending around X86 binaries on the Internet, which is an extremely dangerous thing to do, even if they're encrypted, which is the model that Microsoft has really been pushing. They could contain viruses prior to encryption. The language can be hacked very easily, and it will be hacked. The security problem will come to light eventually because the risk is so high.

So Visual Basic does not really offer a viable solution for network-centric application development, and the world is becoming one giant network. Basically we're all going to become connected by a network, and all the applications are going to have to run across networks. That's why Java's the right choice: Because it's a language that's designed for the '90s and beyond. Visual Basic and C++ and C are languages that were designed in the '70s and '80s, where all the applications that ran, ran on a standalone machine on your desktop.

Leaving Sun: the Java startup

JavaWorld: So you're now leaving Sun to form a startup focused on Java. What prompted the move now?

Polese: There's now a new business unit that's being built up around Java, so lots of support from Sun, from Scott [McNealy] to really build this into a serious business, and that means it's in good hands. I feel confident now that I can leave and the Java machine will keep going, and great things will come from JavaSoft. Java is in good hands.

That, combined with the fact that there were three incredibly talented people who wanted to go off and do something cool with me. So the timing of the fact that all four of us had come to a point in our lives and our careers where this was the right time to do it. And I realize this combination of people and the timing in the market, the need for what we're going to go out and do, which is to build Java applications and to basically deliver on the Java vision and make Java a reality for the average person. Deliver the capabilities of Java to the average person. That market opportunity is massive.

So the combination of the people being there, feeling like I completed what I set out to do, and the market opportunity was just unbeatable. It was very obvious to me that if I walked away from this I'd probably kick myself for the rest of my life. It's something I want to do very very much. There's no guarantee of success in anything, but I think we have a pretty good shot.

JavaWorld: What are your plans with the new company? Without discussing anything specific, can you categorize your new company's efforts? Are you looking at a technology that will be applied by developers? Are you focusing on a single application written in Java that will be used broadly?

Polese: You know, I even feel uncomfortable going into detail there, because people will take what I say and extrapolate and interpolate and interpret.

Learn more about this topic

1 2 Page 2
Page 2 of 2