I was one of the authors of the report in 1995 recommending that IBM change its business strategy and use Java. Customers are moving toward the separation of software from platform and the separation of data from software. IBM's message today is that the operating system wars are irrelevant.... It's all about the Web. Whenever you look at the Internet, the real issues have to do with human beings, not with technology. The really big thing about Java is...there's no longer a link between the software and the platform. XML brought the same sort of philosophy to data...the link between the software and the data is broken now, as well. We have Java support in everything. XML support is going into everything. IBM was an arrogant company at the end of the 80s, and we had a near- death experience. As with many individuals, that...helps you to focus your mind wonderfully.

An Interview with Simon Phipps, IBM's XML and Java Evangelist by Lori Houston Even though he lives in England, IBM's Chief XML and Java Evangelist Simon Phipps is at the center of Big Blue's strategizing with cutting-edge technologies. Yet when he's asked to speak--to deliver, for example, the keynote "Why the Web Works" at O'Reilly's upcoming Conference on Java--he usually doesn't limit himself to talking about IBM strategy and products. It's not that he won't discuss the company's Java-Linux initiatives or IBM's relationship with Sun Microsystems. He simply prefers probing larger Internet trends. As a preface to his keynote address, O'Reilly asked Mr. Phipps to talk about IBM's late-breaking developments with Java and Sun, and the connections IBM sees between Java, XML, and the Web's future. Houston: How did you come to be IBM's XML/Java Evangelist? Phipps: Once upon a time, I worked for Unisys; but I left to join IBM as a product quality assurer. This was in 1990, and I was hired as what polite people call an "agent of change" and impolite people call a troublemaker. Eventually I ended up working on an evaluation of a technology from Sun called Java. I was one of the authors of the report in 1995 recommending that IBM change its business strategy and use Java. That got me into explaining why on earth we had recommended this. Imagine a group of five foreign employees working at a foreign laboratory writing reports to the chief executive at IBM, saying we think IBM should embrace the technology from one of our competitors and change all our existing strategies. That's exactly what we did in 1995. Fortunately, the corporation said "yes" and chose to work with Java. I've been explaining why we did that for the last five years now. Houston: What can you tell us about IBM's recent Java-Linux initiatives with Red Hat, Caldera, and TurboLinux? Phipps: I was very thrilled when IBM announced relationships with RedHat, Caldera, and others. I think that's the beginning of a trend. Looking inside IBM at the moment, we have a very strong emphasis on Linux. We anticipate our customers demanding more and more focus on Linux. Customers are moving toward the separation of software from platform and the separation of data from software. Once you have taken the step of separating your software and your data from the platform and each other, it is no longer such an issue which operating system you use. You can begin to make your operating-system choices on merit and not because you are tied to a proprietary solution. Consequently, IBM has identified Linux as a good solution to our customers' problems because it offers a non-proprietary path towards a stable and functional operating system services layer that complements Java and XML. We have set our direction to make sure that Linux is widely available and is of industry quality. Houston: Will there be more of these IBM alliances with other major Linux vendors? Phipps: Those relationships with well-known Linux distributors are very much about setting the course to have Linux support throughout our product range and across all our hardware. You're bound to see more deals of a similar character coming out as the year goes on. Houston: That makes a lot of sense, but to what extent, if any, do these initiatives represent a "break" with Sun regarding the future of Java? Phipps: I want to be very cautious because, on the whole, I don't like to discuss other companies and their strategies. As it happens, we've got a very close working relationship with Sun. The things we did back in 1995 to adopt Java within IBM were fundamental to Sun's success. Within any relationship, however, you can expect there to be high points and low points. (And you can always expect the low points to be the ones that journalists are interested in!) At the moment, IBM and Sun are having quite a difficult time working through what the next phase of Java will be. Yet we still work very closely together. Back in 1996, we actually opened an IBM branch office next door to Sun on DeAnza Boulevard in Cupertino [California]. It's still there. We still see Sun regularly. It's probably well known that a number of IBM employees have ended up at Sun. What is less well known is the fact that a lot of Sun employees have ended up at IBM. Houston: Why are Java and Linux so strategically important to IBM? Phipps: Do you remember back in the mid-to-late 1990s when there was a lot of talk about which browser was going to be the king, whether it was going to be Internet Explorer or Netscape? IBM took the position back then that the browser was irrelevant. What we're seeing today is a similar war with operating systems. We see Windows 2000 coming out; we see an attempt to defend Solaris; we see a lot of interest in Linux. IBM's message today is that the operating system wars are irrelevant, too. It's all about the Web. The step we've taken in focusing on Linux is more to say, "This is a good operating system, why don't we use it?" rather than about saying, "Let's make an attack in the operating system wars." In the future we'll all look back and wonder what all the fuss was about with Solaris and Windows. In five years most people probably won't know what operating system they're running. They won't care. They won't know which vendor made it, and it won't be an issue anymore than browsers are really an issue today. Houston: Meaning that everything will be Web-based? Phipps: I think that that's the direction we're heading--towards everything being connected. Now, again, you can get philosophical about that phrase "Web-based" and what it's going to mean. In the future, we're going to have an Internet where every device is addressable. You will probably have at least six IP addresses on devices on your body, let alone around your home or around your office. And, in the future we'll see those devices and the ones near them conspiring to solve your problems. Will that be the Web that you're wearing on your body? It will be the Web in the sense that it's connected to the rest of the Internet. And when those devices conspire to solve your problems, they can draw on wide resources. It won't be the Web in the sense of a local network running around you to help you pass incoming messages from your telephone to your PDA [personal digital assistant]. So, when we talk about "Web-based," I think we're going to see our thinking develop in the future a little. It isn't going to be about sitting at a PC and connecting with a Web service. It's going to be about devices conspiring to solve your problems. Simon Phipps will deliver the keynote presentation, "Why the Web Works"at O'Reilly's upcoming Conference on Java, March 27-30. Houston: And since not everyone sees this vision, most companies are not able to develop business strategies around it yet? Phipps: Yes. That doesn't surprise me because this vision of devices solving your problems is really a couple of generations further out in e-business. At IBM, we've thought a lot about anticipating future e-business generations. In other words, first there was the browser generation, and now we're just at the strong point of the consumer e-commerce generation. We're also seeing the growth of the business-to-business e-commerce generation. After that we'll probably see a generation consisting of automated business-to-business transactions, which might be described as "e-marketplace." Beyond that, we'll begin to see devices conspiring to solve problems. Within IBM, we think we're about three to five years into a 20-year change cycle. At the end of that cycle, we'll see a world where devices are conspiring to solve our problems. It would be foolish to spend money trying to create a company that provided that service today because that generation is too far away. Houston: And presumably there are other potential things that could enter into and change the picture. Phipps: Oh, yes. You have to be brave to predict which Internet businesses are actually going to be financially viable in a year's time--let alone what the world is going to look like in twenty years. Personally, I see that many of the issues that need to be solved on the Web are actually matters of human relationships. The human problem is one of choosing whether or not to empower individuals to spend your company's money. Whenever you look at the Internet, the real issues have to do with human beings, not with technology. Houston: Does this speak to why IBM sees Java and Linux as being so strategically important: because these are so critical to future developments with the Web? Phipps: Partly. The reason Java is important touches on another interesting issue here. You have to ask the question: why does the Web work? The reason the Web works is because the consumer and the vendor are able to have a relationship with each other on the Web--without having to be personally linked. That is, the Web has minimized dependencies. The computer industry on its own--not vendors--has identified technologies that allow you to minimize dependencies between solutions, and minimize dependencies between solutions and platforms. There are really four areas where those technologies apply: 1) computer- to-computer networking, 2) delivering solutions to put them in front of users, 3) building software solutions, and 4) representing the data that the solution is using. About two years ago, that picture became complete. We had TCP/IP for the networking, we had the use of Web servers to get content in front of people, and we had Java, and specifically Java components, as the way that you write software. Because the really big thing about Java is not the programming language. People misunderstand this. It is the fact that you end up with portable binary programs that don't care what computer they run on. And because you have that, there's no longer a link between the software and the platform. Then, XML brought the same sort of philosophy to data: the idea of using the field names of your data as the delimiters within the data. Because the structure and meaning of the data are in the same place, you no longer need to resort back to the software to understand the meaning. So, the link, or dependency, between the software and the data is broken now, as well. Taken all together, those four technologies tend to break dependencies. With TCP/IP you're no longer dependent on a proprietary networking architecture--if anyone can even remember the days of proprietary networking architectures. With Java, there isn't a dependency between the software design and the software deployment. And with XML, there isn't a dependency between the software and the data. I can publish data in XML and you can consume it; but we don't have to use the same software. Houston: Do you expect to see the use of Java and XML continue to grow in server-side applications? Phipps: Yes. IBM identified that our business was to help our e-business customers succeed, from software to servers to services. Consequently, we had to go in and re-tool everything so it supports those standards: TCP/IP, Web servers, Java, and XML. So that's what we've done. We've completely changed all of our products. We have Java support in everything. XML support is going into everything. Everything is designed to work within an e-business rather than a stand-alone or proprietary environment. That's why Java is so important to IBM: It's the vehicle with which we can still write the software with the power to solve problems--without making our customers become dependent on us--or anybody else. The whole point of this strategy is really rather obvious. I'm still looking for the source of the quotation that an American president once said: "It is easiest to persuade a man of that which is obvious." We're not telling the industry to use TCP/IP, Web servers, Java, and XML. We're observing they do use these technologies, that they do fit together, and that there is one unifying philosophy underlying them: making things work in a connected world. And we're just going out and doing it. I think IBM is actually a little different because many of our competitors are still stuck in the world of proprietary technologies and focusing on operating systems or just one standard and wanting to lock in their customers. People can't afford to do that sort of thing any longer. Houston: Considering IBM's beginnings, that's quite a shift. Phipps: You do have to remember that IBM was founded by Herman Hollerith as the Tabulating Machine Company, and remember we started out doing punch cards for the US census. We had a phase when we made typewriters and adding machines. We made door locks and we used to make clocks. We invented electronic storage media like hard disks, and we invented the idea of the virtual machine. IBM is probably the leading company doing things like MPEG at the moment. We chair the MPEG-7 committee. So we are used to evolving, following the trend from the front. To be frank, I think IBM was an arrogant company at the end of the 80s, and we had a near-death experience. As with many individuals, that sort of experience helps you to focus your mind wonderfully. The IBM I work for now is not the same IBM that I joined in 1991. What you see now is an IBM that values innovation and is trying to surface and expose that innovation, and is trying to do so in a way that embraces the idea of the community and that values standards. Houston: Do you have anything else that you would like to add? Phipps: My main emphasis is always to say that it's not about technology, it's about people. If you want to look at how everything works on the Web, if you want to understand the future of where the Web is going, you need to understand people and how they relate individually and in businesses. Once you understand that, you can make a lot more sense of where we're going next. If you enjoyed this interview with Simon Phipps, check out his keynote presentation, "Why the Web Works" at O'Reilly's upcoming Conference on Java.