|
Is Sun taking the initiative on openness and community, or is stepping off into the future by taking the middle course just being reactive?
June 30, 2006
URL:http://www.drdobbs.com/jvm/sun-java-and-the-middle-course/189401579
Onno Kluyt leans forward intently. He's doing interviews all this week, and here he is leaning across yet another generic table in yet another interchangeable room at Moscone South with an assiduous PR person at his elbow and a skeptical journalist in his crosshairs. Kluyt's job at JavaOne is to talk up community, to push this message of openness relentlessly to the press and to developers, and if he were anything less than earnest about the task, it would be painful to watch.
But there is no mistaking Kluyt's commitment. He's the chair of Sun's Java Community Process (JCP) program (jcp.org), and Sun's commitment to community is not just his message, it's his mission. "JCP defines what Java is," he tells me. "It's where to stay abreast of developments in Java, even if you're not active in the JCP. It's the one place where Sun can't ignore you."
This idea of community is not just some public relations gimmick Sun dreamed up last week. It's not unreasonable to say that Sun has always been about openness. "We were the Red Hat of Berkeley UNIX back in 1982," Scott McNealy has said. Andy Bechtolscheim's original workstation that convinced Vinod Khosla to start Sun was a model of building with off-the-shelf parts. With Java, "the open story has been going on for 11 years," Community Marketing Manager Rich Sands tells me over another of those tables at Moscone. The arrival of Java also brought increased emphasis on another capital-C theme about which Sun is obsessive—compatibility, encapsulated in the slogan "Write once, run anywhere."
But Sun is currently facing challenges on many fronts. Open Source projects like Eclipse challenge its development tools and challenge its commitment to openness. There's a challenge in meeting demand for programmer productivity and speed of development that is being thrown at Sun by focused development environments and tools like Ruby on Rails. Microsoft is a perpetual challenge, with dot-net now the elephant in the room. There's profitability, of course: The bottom line is challenging Sun now. And, arising out of all these challenges, there's the challenge of a dubious image.
Against that backdrop, seismic changes are rippling through Sun management. After 22 years, CEO Scott McNealy has passed the reins of the chariot to Jonathan Schwartz, who may be the most open CEO in the Fortune 500: At least he's the only one with his own blog. Schwartz, the buzz has it, is making Sun interesting again. Rich Green, who was once Sun's chief Java advocate and played key roles in Java Studio Creator, JVM, and Sun's battle with Microsoft over Java, has returned as EVP/Sun Software. Last year, cofounder Andy Bechtolscheim returned in an acquisition after a nine-year absence. Other returnees include Peter Ulander and Karen Tegan-Padir.
Largely because of these changes, BusinessWeek recently picked Java, familiar old Java, as a key technology to watch. That actually makes sense: With the exception of Bechtolscheim, most of these people are software focused, and most are focused on Java. It is reasonable to expect a significant shift in emphasis at Sun, placing (even) more emphasis on software, on Java, and on the themes of openness, community, and compatibility. "Software is in the ascendant at Sun Microsystems," Gavin Clarke of The Register keeps telling us. At this point I think we can safely add, "No duh."
When Apollo passed the reins of the chariot of the sun to young Phaeton, the godling's maiden flight didn't go well. Change always carries risk, but not changing with the times carries the certainty of death. For a high-tech company, it's not a question of whether you change with the times, but how.
Oops. I really didn't mean to steal the best line from JavaOne. You know:
Schwartz, standing in front of 14,000 developers: "Will Sun open source Java?"
Green, looking uncomfortable: "It is not a question of whether, but...of how."
How, indeed.
If I were Rich Green, I'd prefer this not to be the question on everyone's lips, even if I knew the answer, because it masks the consistent commitment to openness at Sun. But there's your CEO front-burnering the question in front of 14,000 developers, so it's got to be addressed. Just Sun being open about openness, you might say.
So if those are the accepted terms for the discussion, and if it's not a question of whether but how, we'll bite: How?
My hope is that Schwartz and Green think they know the answer or they are prepared to reexamine the issue with fully open eyes, considering, as Schwartz says they are, even a GPL license. I wonder. Can they actually open source Java, just like that, no obfuscation or equivocation or inventing new forms of openness, but just do it?
And what does Sun get out of it? And what does Sun risk? And is it too late?
What Sun gets, Schwartz says, is increased profits, at least eventually. Sun made gains in revenues and marketshare in servers in the first quarter of this year, and Schwartz says it was partly due to the open sourcing of Solaris (www.opensolaris.org/os/). And Solaris points the way for Java, he says. The way is not one of immediate payback, though: that comes from selling services to the market you create with the open sourcing of technology. The Register: "That, according to Schwartz, means revenue becomes a lagging indicator of the adoption of Sun's developer platforms...it's OK to lose money up front, because revenue will flow further down the line."
Revenue as a lagging indicator of success: I've worked for companies with that business model. I hope it works better for Sun.
JavaOne, in May of this year, was the Sun-watchers' big opportunity to take readings of the magnetosphere. Are Sun's poles flipping? Will south become north, or maybe south-southwest? Sun obliged us with many product and technology announcements. These weren't all about open source and community, although Sun sort of tried to cast as many of them as possible in that light. Among other things, Sun announced:
I leave it to Eric Bruno to dig into the details of these announcements on his blog (www.ddj.com/blog/javablog/). But how do these announcements speak to Sun's challenges?
The open-source initiatives directly address the open-source challenge, and follow on logically from the Solaris move. In the year since Solaris went open source, more than 5 million copies have been downloaded, and Oracle moved from Linux to Solaris as its preferred base OS, a coup. With the specific Java technologies that Sun has open sourced, the company does seem to be shifting its poles a little, but open sourcing Java itself would be a major flip.
The Linux license was developed in consultation with Linux vendors, but it has nothing directly to do with open sourcing Sun technology, as everyone from Richard Stallman to the folks at Redhat has pointed out.
NetBeans is in an uphill fight against Eclipse, which has many more plug-ins and users. Java creator James Gosling says there's no way Sun will join Eclipse, so Sun seems prepared to fight this one out from an underdog position.
As for the agility/productivity challenge, Sun has seen that, as the cost of tools comes down and all IDEs try to deliver the core capabilities of their competitors, programmer productivity becomes a more important differentiator. And Sun's tools have been branded with a C, not for community or compatibility, but for complexity. Green indicates that Sun plans to meet the challenge of tools like Rails by making its own tools more productive but also by making it easier for developers to use Ruby and other scripting languages in Java projects.
The bottom-line challenge will get tougher, not easier, if Sun gives away or gives up control of more of its software. According to Schwartz, services will take up the slack, but he's not terribly precise about what these services are. And what if those services are adopted as enthusiastically as, say, NetBeans? A review of Java IDEs in JavaPro magazine judged NetBeans to be at least comparable to Eclipse, but pointed out that it "lacks the community and developer enthusiasm of Eclipse." Ouch. No wonder Sun is obsessing about community.
But is Sun taking the initiative on openness and community, or just being reactive?
"We've learned a lot about how communities connect," Kluyt tells me. He paints a picture of a JCP with more power in the hands of individual programmers, one less dominated by corporations. What he says next is clearly a rehearsed slogan, but he seems to mean it: "JCP is about developing standards by developers for developers."
"It's been an education both ways," Rich Sands adds. Today, "a formal part of being a Sun engineer is interfacing. Integrating bug fixes, blogging. This the model of the future of software development."
Sun has faced down daunting challenges in the past by embracing openness. It hopes to do so again, and JCP is central to that effort. Sun wants there to be no confusion on this point—it wants everyone to consider joining the JCP.
But since Schwartz encouraged the already incessant drumbeat for open sourcing Java, the effectiveness of its JCP efforts may all come down to what it does regarding this delicate issue.
Phaeton needed to steer a middle course between driving the chariot too close to the earth and too high in the sky. Sun is looking for the middle course that gives them just enough control and openness that balances compatibility and community. There's a lot riding on its finding that middle course.
Terms of Service | Privacy Statement | Copyright © 2024 UBM Tech, All rights reserved.