JAVAONE: Sun’s second act?

When F. Scott Fitzgerald observed in his (unfinished) final novel, The Last Tycoon, that there are no second acts in American lives, he was himself struggling with the pitfalls of a failed reinvention. Famous at 23, washed up at 40, dead at 44, Fitzgerald’s life epitomized artistic burnout and provides a cautionary tale that still haunts the American zeitgeist these many generations later. For Fitzgerald-the-screenwriter, life was never as bold or as sweet as it was during the height of his novelist adventures. The question that kept occurring to me at this year’s JavaOne was rooted in Fitzgerald’s last season in Hollywood: are second acts ever performed by US companies? Because, for Sun Microsystems, the curtain on the first act has most assuredly fallen.

End of Act One

I entered Moscone Center for the 2004 JavaOne Conference rather skeptical. My attendance record claims all but two JavaOne conferences: I missed the first, taking my first Java-related assignment at Sun a few months after that very first JavaOne. I also missed the 2003 conference, having left Sun in November 2002 (after nearly 9 years at the firm and 20 in software, I needed time away). But I have attended all other instantiations of JavaOne at Moscone Center. This year, the first as an independent consultant, I did not attend to serve, but to observe.

Having been inside, as it were, I have witnessed first-hand the “behind-the curtain” angst that necessarily occurs prior to any major event. That coupled with the fall from grace that has defined Sun since (coincidently) Bush was elected gave me considerable fodder for a skeptic’s fare. And if signs of decay were all I wished to observe, there were ample to sample:

• Less attendance: I estimate there were (maybe) half as many attendees as there were at JavaOne’s peak—the ’99 and ’00 events burst the doors off Moscone. Although McNealy’s keynote suggested that 14,000 attended this year, I doubt the accuracy of that estimate. If there were 9,000, I would be surprised.
• Fewer give-aways: No iButton-powered JavaRing. No PDA. No backpack on wheels. The attendees got a small shoulder bag, and that was about it. Once known for cool chachkas, JavaOne has slipped several notches in the give-away department.
• Fewer supporting partners: The pavilion floor was nowhere near filled with startup Java ventures as compared to past years.
• Fewer industry-molding revelations: A dribble of open source announcements (Java 3D and Project Looking Glass) were all that Sun could muster.

I heard the fat lady warming up.

I ran into a number of old friends and acquaintances from Sun, many of whom are still with the company. The giddy, “glad-to-be-here” temperaments of the dot-com heyday were entirely absent from Sun personnel, replaced by somber, reality-tested mantles of “survivor,” having witnessed a bevy of now-ritual RIF events. Three years of profitless results must test the faith of even the most ardent believer.

It was interesting to see the pavilion floor layout, with Sun booths dead center and Sun’s now red-shirted support staff tightly compressed in a “circle-the-wagons” configuration. The space between Sun booths was literally a third that of all other pavilion participants, making it at least uncomfortable to enter the inner sanctums of Sun’s offerings—clearly an operational choice made with ironic strategic overtones. Avoidance was the only rational choice for attendees, which, I suspect, most accomplished. It was simply too difficult to get inside the Sun displays to speak with a Sun booth representative. The self-defeating subtleties didn’t stop there…

Bill Joy was absent. Having left Sun himself, there was probably no reason for the former chief scientist and Sun founder to attend. And while his presence was never assured at prior shows, he did manage to attend a few previous JavaOne conferences. His presence this year might have been viewed as positive for his struggling former company; his absence assured more questions.

Bill Gates didn’t make it either, but his shadow was evident throughout the event, one of those elephants in the room that no one would acknowledge. The sound of Scott McNealy biting his tongue echoed loudly after his keynote. Now that Microsoft bashing is no longer his main shtick, the middle-aged CEO was obviously missing the absence of his decade-tested quip base. I could hear him getting close to a Redmond slam and then watched as he repressed what must by now be instinctive. The only time he let it slip was during his wrap-up, a “Where’s the Outrage” junket, when McNealy cited viruses (as in Microsoft viruses) as a source of much global duress and yet little apparent institutional (read governmental) concern. “There are only two developer communities left on the planet,” he had said earlier. Bridging the two (Java and .Net) is the essence of the 10-year collaborative agreement between the two firms. Does that mean that Java and .Net will interoperate? Probably, but I wouldn’t hold my breath as to the sanctity of the Java community or the continuing cleanliness of the code base over the next few years, Sun’s “benevolent stewardship” notwithstanding.

Also MIA was John Gage. The venerable Gage, Java’s chief chaplain, had other duties to attend to. Gage, whose steadfast, father-knows-best personae has grounded the event since it began, had better things to do, leaving a relatively young, wannabe comic (short of writing staff, it would seem) to emcee Sun’s flagship event.

It’s going to be another long year for SUNW. Alas, end of Act One. Curtain down.

The Big Question

What has set Sun apart from so many other companies is a certain philosophical thread, an aberration of corporate thinking, which is almost anti-capitalistic at times. Consider Joy’s Law, an intellectual break-through that is the essence of the open source movement: “Innovation happens elsewhere.” The long version of Bill Joy’s contribution to corporate IP acknowledges that no matter who you are, the smartest people in the world do not work for you. Innovation will inevitably occur in other locales, with other teams, by other geniuses in other companies. As such, if one would benefit from that inevitable innovation, one must embrace a strategy that would leverage such occurrences: hence, open source.

Sun has a long history of contributing to and supporting open source and community source kinds of efforts. Despite recent wrangling by IBM (and others) that Sun should fully open source Java (the Big Question), Sun’s record with respect to the opening of source bases is far superior to that of any firm…ever. The debate on the last day of JavaOne, in which Sun’s James Gosling and Rob Gingell joined Brian Behlendorf (CollabNet), James Governor (RedMonk), Justin Shaffer (MLB Advanced Media), Rod Smith (IBM), and Lawrence Lessig (Stanford University) with moderator Tim O’Reilly (O’Reilly Media) in a discussion of open source Java and its merits versus the continuation of the benevolent stewardship role that McNealy articulated (“somebody’s got to be in charge or nobody is”), was instructive but unclear. It is somewhat ironic that Sun is on the defensive in this discussion, in light of the firm’s history—it’s like John Kerry defending his Vietnam War metals against a National-Guard-absent George W. Bush. But clearly, a more open Java is on the table and likely to be a contentious issue until it isn’t.

Does an open source Java even matter? What does “control” of Java mean? According to Scott Stark, JBoss CTO, open-sourcing the Java runtime engine is far less important than full transparency of the TCK. “It [the TCK] should be open sourced to help all of the J2EE community,” he said during our discussion. “We need to be able to understand what it means to successfully navigate the certification process. That’s what’s important.”

The panel too was unclear as to what exactly is meant by open source—what part of the code? Is there a problem, or is it simply perception? What’s wrong with the JCP (Java Community Process)? Are these licensing issues or cultural issues?

The extent to which the JCP may stifle innovation may be the crux of the matter. To my way of thinking, it is yet unclear that any other community approach would be superior to preserving the benefits of the Java platform without risking divergence and the horrors of creeping incompatibilities, the raison d’etre for Java from the outset. I recall the warning from my own book, Network Distributed Computing: Fitscapes and Fallacies, which I was promoting at JavaOne, in terms of evolution and the notion of fitscapes: innovation cannot proceed faster than the containing fitscape can test the novelty without risking systemic failure. As such, the JCP could be the best we can hope for in light of Joy’s Law and the tendency of change to accelerate in these interesting times.

Act Two?

So now, whence goeth Sun? Years of loss can be brutal and become habitual. Clearly, if Sun continues on its current death-spiral, it will inevitably go the way of all other companies whose time has come and gone. The Java community, the JCP, the evolution of the platform…all could be moot if Sun insists on maintaining it’s benevolent steward role but fails to deliver substantial financial results. Java’s emergence represented the dramatic peak period of Sun’s financial performance, which has been followed by an equally dramatic decline.

From 1988, when Sun’s revenue first passed billion, to fiscal year 2001, when Sun’s Y2K-inflated revenue exceeded 8 billion, the company seemed to be headed for the same meteoric rise as arch-competitor Microsoft. But the combination of the dot-com implosion (the bubble of which Sun helped fuel with “We’re the dot in dot-com” marketing), the 2000 election uncertainty, and ultimately the destruction of the twin towers on September 11, 2001 (Sun had two floors in the south tower of the World Trade Center), top line growth for the firm did an abrupt about-face and has never looked back. Fiscal year 2003 yielded revenue results less than those of FY99 for the company; FY04 news should be public any day now, and it will be interesting to see if Sun can reverse a three-year decline. But if the first three quarters of Sun’s most recent fiscal year are any indication, the chances of Sun demonstrating revenue growth in FY04 are slim indeed. In all likelihood, the decline will continue.

The ’90s were heady times for Sun. The rise of Java, the burgeoning Internet (which Sun helped pioneer), the emergence of “geek chic,” the network-age technology acceleration which Sun explained, applauded, and profited from, even while proffering a “stuff-should-be-free” ethic…all this and the growth of Java gave rise to a bloated firm addicted to its own press. But the reality of the twenty-first century has changed all that; a shrinking, RIF-battered Sun circles the wagons, speaks softly to itself in fear and denial, and waits for a return to remembered better days, holding tightly to the illusion that history has thrust a benevolent steward role upon them, one which they cannot shirk without risking a decline into the oblivion of Tower-of-Babel incompatibility.

So I’m buying SUNW when the markets open this week. “What?” you may ask, “Are you insane? Why, after all the seemingly bad news you’ve reported from JavaOne, would you risk buying SUNW today?” The answer is simple: at the proverbial end of the day, I think there’s hope for Sun. And I came to that conclusion during the first keynote.

As stated above, I entered Moscone a skeptic, ready to record all the evidence of decay and denial, which I knew would be pervasive at the conference this year. And yes, I found evidence of decline, some of which I’ve reported here. But I also found reason for hope. If there’s hope for Sun, it’s with its new COO and president, Jonathan Schwartz, who gave the keynote on day one of the event. As difficult as it may be, I think he’s got the right stuff to make the necessary changes within Sun to turn things around, which is ultimately going to be necessary if Java itself is to continue to be relevant.

Schwartz is not a cheerleader. Nor is he a general. I’ve only met him once, during a Sun internal event a couple of years back. But he leaves a lasting impression, one that translates well from one-on-one encounters to event-sized audiences. He’s very bright, well spoken, thoughtful, and perhaps even wise. And he seems to understand twenty-first century economics better than anyone I’ve met in corporate Globmerica to date. Listening to him can be confusing if you take him at face value. But this year, at JavaOne, I heard beneath the level at which he was speaking (which I think was his intent), and got the message. It’s about ring tones…

See, it’s about ring tones as metaphor. Downloadable ring tones. Downloadable ring tones with a global market of .5 billion today. That’s a lot of toast in a toaster-net world. Schwartz gets that future services will be like ring tones, downloaded to subscription-based networked devices that are essentially free from the consumer’s perspective. In Schwartz’s world, hardware, in agreement with Bill Gates, will be free; a component of a profitable-NPV subscription-based service that will be fee-based and extensible. In Schwartz’s Sun, reliance on thick-margined servers goes away (as has the market) in favor of recurring revenue, which, by coincidence, represents about of third of Sun’s current top line. If Sun gets that vision and can manage to manage customers’ needs instead of a circle-the-wagons insular view (which may also be a tradition at Sun), it is possible that an investment in SUNW will be profitable in the next 18 months, which means Java will last. If not…well, like so many other great companies, Sun will effectively go away, absorbed by Fujitsu or its moral equivalent, and Java will fracture, bifurcate, and, in the process, evolve, away from universal compatibility, to be sure.

Sun claims 4 million developers embracing Java. Such claims are difficult to substantiate, but it is probably true that the Java community has continued to grow, despite the decline of its benevolent steward. If Sun has an Act Two, then perhaps the best thing for the Java community is for that stewardship to continue. But whether the play continues is up to Sun. At this juncture, the air is ripe with the smell of decay. If there’s hope for Sun, it’s in Schwartz. Follow that thread, and the future of Java will become increasingly clear.

Miscellaneous notes

The most interesting products/vendors on display at JavaOne:

• 2AB: Trusted solutions for distributed business
• Enerjy: Cool Java tools
• Skyway Software: More cool Java tools
• Recursion Software: Multiplatform developer tools
• Handango: Wireless/handheld software publisher
• Sonic Software: Enterprise Service Bus purveyor

The most interesting press conference was an SOA (service-oriented architecture) panel held on day two—interesting because of the lack of clear definition as to exactly what SOA meant, interesting because it represents a clear growth segment in enterprise software development, and interesting because I ran into Miko Matsaumura (The Middleware Company), aboriginal Java evangelist.

The most interesting demo was Simon Ritter’s during a James Gosling segment on the final day. Ritter, a dear friend and former cohort, presented a prototype augmented reality demo that I suspect will mature over the next several months. While a bit rough, the possibilities of augmented reality were actually demonstrated, something I have read about and imagined but have never actually seen.

The most interesting technologies to watch in the next 12 months: network-served biometrics, anything mobile, Java-.Net bridges, and SOA stuff in general. Oh, and ring tones.

Disclaimer: After nine years with Sun, the last six of which were spent serving as a Java Technology Evangelist, it is likely that I retain a substantial pro-Java bias, despite two years Kool-Aid free.