“Oh crap,” thought James Gosling. The Sun Microsystems vice-president was walking down a dark alley in a seedy part of Paris, and a shady-looking man was on his tail. Gosling headed for a street lamp. Once there, he turned and saw his pursuer approach. “Oh, you’re James Gosling,” said the gushing man. “Can I have your autograph?”
The Calgary native may not be a household name, but among geeks who program in Java, a computer language Gosling co-created at Santa Clara, Calif.â??based Sun in 1995, he’s someone to be worshipped. Some fans have trouble speaking in his presence. Others tug at his clothing to convince themselves he’s real. And among the 1,000 or so members of the Java Community Process program who toil to improve the language, he’s considered a god.
This awe of Gosling is all due to the wild success of Java, which the 53-year-old Canadian designed to work on a wide range of hardware platforms. Bits of its code are embedded in countless web pages in the form of tiny programs called “applets,” which do everything from updating sports scores in real time to calculating mortgage payments. Many cellphone games are Java-based. Banks rely on the language to process financial transactions. NASA uses it to select landing sites on Mars. Sun, which generates about US$13 billion in annual revenue by selling servers, storage devices and support services, estimates Java runs on billions of devices around the world.
But Gosling appears ambivalent about his creation. Sure, Java has allowed Gosling to feed his insatiable curiosity. He’s been able to visit famous research facilities such as CERN, the particle physics lab in Geneva, and met fascinating people, including Father Michael Gardener, a priest who walked across Nunavut opening schools. (They met when both were awarded the Order of Canada in 2007.) But Gosling is uncomfortable being called the inventor of Java, since many people have contributed to the language. And during presentations outside the company, he sometimes feels like Vanna White, someone used to jazz up the show. He’s constantly in demand, which often prevents him from doing what he truly loves: writing software, an activity he finds so engrossing he can do it straight through the night. “In some sense, the success of Java was one of the worst things that’s ever happened to me,” Gosling says.
These days, Gosling’s role at Sun, which he joined in 1984, is to be a rabble-rouser for Java. At Sun’s annual JavaOne conference in May, Gosling loaded rolled-up T-shirts featuring one of his designs — an adaptation of Edward Hopper’s Nighthawks painting, but showing Sun mascots seated in the diner — into a giant slingshot and launched them into the audience. He also hosted a showcase of Java applications, including Tommy Jr., a Toyota Scion that drives itself. But the event happened in the wake of disappointing quarterly results by Sun, which is raising questions about the company’s investment in Java and, as a result, Gosling’s importance to the company.
Gosling describes Sun as “a family of people just kind of doing crazy things.” In the early ’90s, he led a group charged with the ambitious task of developing a programming language for the next generation of computing. The team settled into an off-site office and — after 18 long months — emerged with “Oak,” named after a tree outside Gosling’s office. Sun’s execs thought the device-independent language would be ideal for personal digital devices and digital television equipment, but those markets failed to go mainstream. A demoralized Gosling considered leaving Sun. But Bill Joy, one of Sun’s co-founders and then-vice-president of research and development, realized Oak’s anti-virus properties would make it especially useful on a new technology platform called the World Wide Web. Joy and Gosling convinced Netscape to incorporate Sun’s new programming language — renamed Java — into its popular browser. Sun released Java’s code to encourage programmers to develop applications, and, soon after, Java took off.
But Java’s value to Sun during the ’90s helps explain the questionable link between the programming language and the company’s bottom line today. When Java was a fresh idea, Sun’s sales force could leverage the curiosity of CIOs to pitch them on the company’s servers and support services. Now that buzz from Java has worn off. It’s become so common that it’s not a differentiator that Sun can use against its competition, which includes giants IBM and Hewlett-Packard. Java works as well on Sun servers as it does on IBM. Of course, the language would never have been so successful without this feature. But Dana Gardner, a principal analyst at Interarbor Solutions, a tech consultant in Gilford, N.H., says he has yet to see evidence that Java’s popularity has directly translated into Sun’s financial performance.
Java clearly has an image problem. For one thing, Sun gives away its programs to run and develop Java applications, as well as OpenOffice (an alternative to Microsoft Office) and OpenSolaris (an operating system). With Sun’s US$1-billion acquisition of MySQL earlier this year, the company added a popular database to its open-source arsenal. Sun also provides the programming code of its software. This encourages adoption and entices people outside of Sun to contribute to improvements. “One of the reasons Java has really strong security is tons of people have pored over the source code looking for problems,” Gosling says.
The computer language also works behind the scenes. “Most people are using Java apps every day, all the time. They just don’t know it, which drives our marketing people nuts,” admits Gosling. Nevertheless, Jonathan Schwartz, Sun’s CEO and president, sees tremendous value in tying the Java brand more closely to his company. Last year, the ponytailed exec changed Sun’s stock ticker on the Nasdaq to JAVA from SUNW. “The number of people who know Java swamps the number of people who know Sun,” he said in his blog about the change. As for the value of computer language to the company, he wrote, “Java means limitless opportunity — for our software, systems, storage, service and microelectronics businesses.”
Others aren’t so sure. Paul Massie, the senior director of information technology at Santa Claraâ??based Genesis Microchip, whose display image processors are used in televisions and computer monitors, boils Schwartz’s strategy down to one of getting “the Sun brand everywhere and the money will follow.” He quickly adds, “I think he’s seriously mistaken.” Massie says Schwartz hasn’t demonstrated — or properly explained — how Sun can make money from the open-source model, holding up the company’s recent financial results as proof. He doesn’t buy the fact that having experience with Java or Solaris persuades people to buy hardware or services from Sun. While he recently bought some Sun servers, it wasn’t because he has experience using Java. “For US$20,000, it was simply a really good price-performance ratio at the time,” Massie says.
Schwartz insists an open-source approach to Sun software is the way to capture hardware and service sales. And now he has a new tool to promote. Sun’s latest much-hyped open-source offering is called JavaFX. It’s a programming tool to build “rich Internet applications.” Some of its features include the ability for users to drag applications from their web browsers permanently onto their desktops. (A demonstration of this elicited whistles and applause at JavaOne.) Its applications will be able to show high-quality audio and video. Although not in the initial release this fall, JavaFX will eventually allow software developers to monitor how people are using their own applications, such as which functions are most popular.
Following the JavaOne announcements on JavaFX, shareholders questioned the value of the initiative to the company’s bottom line. As one person put it, “OK, so how does Sun make money and add shareholder value?” Techies may love Sun, but investors are not so happy with its performance. Its stock has fallen 24% following Sun’s most recent quarterly results. Management had promised mid-single-digit sales growth, but reported a 0.5% decline. After six consecutive quarters in the black, Sun was once again in the red, with a US$34-million loss. Management revised its operating margin target to 7% from 10% for the end of fiscal 2009, which will be helped along by the elimination of up to 2,500 jobs by Q2 next year. Schwartz blamed the loss on the slowing U.S. economy, which forced American customers to postpone large orders. Shareholders vented their anger on Schwartz’s blog: “Don’t you feel shame?” wrote one, while another ever so eloquently typed, “You suck J.S.” Morgan Stanley analysts in a research note said, “With the restructuring story derailed and no real catalysts for revenue growth this year, JAVA is likely dead money for longer than we expected.” Gosling, for one, was annoyed by the reaction. “If you looked at how the company did around the world, almost everywhere we did really well,” he says.
If the company continues to lose money, expect changes. Sun may be forced to significantly cut its research and development budget, which hovers around US$2 billion per year, about the same as it was in the heady dot-com era. That means making some tough decisions about who gets what. Sun could cut investment into Sparc, the company’s proprietary microprocessor chip, walking away from a no-win fight with Intel, which boasts an R&D budget nearly three times as large as Sun’s. That might relegate Sun to competing in the low-margin commodity server market, but it could return to its roots as a tweaker of off-the-shelf technology. The company followed this strategy in the ’80s, having tremendous success in the market for workstations, powerful PCs used by scientists, professionals and engineers. Back then, co-founder Scott McNealy said, “What’s proprietary to us is the art of making a Ferrari out of spare parts.”
Sun could also cut its Java budget. Without promoting Java, Sun certainly loses a way of getting its name out there. But it could consider doing that the old-fashioned way: advertising. Sun spent only US$25 million on ads last year, which is dwarfed by its competitors: IBM spent US$1.2 billion, and Intel spent US$1.9 billion.
Gutting Sun’s Java budget would also likely mean a different role for Gosling. He could quietly spend his days writing code for Java as a member of the open-source community. It’s not hard to imagine him enjoying that, nor is it difficult to picture Gosling pursuing new intellectual pursuits. When asked if he’d like to end his career with Sun, Gosling chuckles and says, “It’s hard to tell, but I have a long list of things I want to do that aren’t Sun-related.” If another programming language supplanted Java, that wouldn’t necessarily bother him. “If someone came up with a language that’s actually kind of cool, I’d run over there and check it out.”