Report from JavaOne: Participants speak their mind

Java developers are not shy. I know. I spent four days talking with them about everything from the state of Java to the genuine nature of James Gosling (and the not-so-genuine nature of Alan Baratz) to the fact that Java is attracting too many “suits” for developers’ tastes. And I wasn’t even asking those questions! Lips started moving the minute the JavaWorld logo embroidered on my jacket was spotted. The “Oh, you’re from JavaWorld?” line sounded suspiciously like the one uttered by John Lovitz’s chronic-liar on Saturday Night Live. I kept waiting for, “Java? Yeah, I’m a — uh — developer, a Java developer, that’s right.”

Keynotes get lips moving

Early each morning of the conference I headed down to Moscone Center to hear pearls of Java wisdom delivered by one Sun/JavaSoft official or another (see Resources for links to articles about each keynote along with the actual keynote transcripts and slides). Clearly, the majority of conference goers had the same idea, as seating for the kickoff events was scarce (at least for those first two days).

Gosling, Baratz strike a different chord with listeners

In his kickoff keynote, James Gosling, Sun Fellow and Java language creator, ran down a series of vital statistics for Java, including the number of JDK downloads, the introduction and subsequent success of the JavaBeans (and Enterprise JavaBeans) devkit, and continued developments on the security front. He also addressed the major concerns surrounding Java today — stability and performance.

Not surprisingly, Gosling’s easy-going manner brought high praise from his devotees. “James resonates with us because he’s one of us,” said UMass Computer Science student and part-time Java developer John Oleckman. “He bypasses the marketing hype and goes straight to what matters most to developers.”

Noting his impression of Gosling from this and other encounters, Mathematics professor and Java developer Daniel Steinberg said speaking to Gosling is “like having a casual conversation with a colleague in the Math lounge.”

True to form, Gosling humbly acknowledged those who have made his creation such a huge success, noting that interaction with the Java community is “an essential part of how we run Java.”

When asked which development tool he prefers, Gosling’s response, “vi,” had folks in my near vicinity twittering with excitement. “I use vi, too!” boasted the woman seated behind me. “Just give me a command line and the JDK. That’s all I need,” countered her companion.

It’s hard to follow such a warm reception, particularly when it’s combined with a number of interesting demos, as Dr. Alan Baratz, president of JavaSoft, found out with his own keynote speech, in which he described the next phase of growth for the Java technology platform.

After his 1998 predictions for the high-tech industry fell painfully flat, he was unable to get the momentum moving in his direction. A noticeable number of people excused themselves from the auditorium.

Perhaps the emphasis Dr. Baratz placed on statistics and products didn’t sit well with the audience — even though it’s those statistics and products that illustrate Java’s growth in the industry. In addition to these topics, Baratz covered the worlds of enterprise and consumer devices, noting the opportunities Sun has capitalized on in both arenas.

I found out later that unlike Gosling, many developers don’t see Baratz as a peer. “He’s a suit — even in jeans, he’s still a suit,” lamented Nasran Sudithara, a Java networking specialist from India. “I get the feeling he’s more about money and politics than developers.”

Laughs and sincerity provide strong combo for Sun officials

Day 2’s keynoters, Ed Zander and Scott McNealy, COO and CEO of Sun, respectively, also spoke to a packed house and touted Java’s growth in the enterprise and consumer arenas (newly announced licensees included Sony and Ericsson). Zander talked about a number of ways Java has moved into the real world, including a very personal enterprise triumph for Java within Sun’s own walls — Sun.Net — which represents the largest deployment of Java in the enterprise.

McNealy, of course, took a more lighthearted approach, complete with the now-expected “Top Ten” list. (The only surprise: This time his list didn’t include even a single jab at Microsoft.) He spoke about the three computing applicability models used in the industry to date (Windows, Unix, and now Java) and pointed out why Java, which uses an open, collaborative strategy will win the battle:

When we first launched Java, no one was thinking about using the Java platform for the largest database warehouses and supercomputers. Probably the most amazing thing is how rich the APIs have gotten in less than three years. Microsoft has been working for 20 years on [its] API set. There are not many applications you can write in Windows but not in Java. You cannot write an application in Windows for your ring. You don’t run Windows applications on your mainframe, your supercomputer, and your scalable reliable devices. And you don’t put Windows on your consumer devices because consumers don’t like to reboot.

Although he portrays himself as a “golf major” and tends to give the impression that the label is not so far from true, he commands much respect for his single-mindedness when it comes to networks in general and Java, specifically. “His ‘Top Ten’ list always gets me going, but it’s his sincerity about Java that makes me really listen to what he has to say,” commented Sharon Steed, a programmer with Netsmart Technologies in Toronto, Canada. “If they [Sun] don’t come through on the promises they made yesterday [referring to Dr. Baratz’s promise that Sun will be focusing on performance, stability, and compatibility], you can bet that ‘scott-dot-mcnealy-at-sun-dot-com’ will be hearing from me!”

Panel discussion goes over the top — of listeners’ heads

The third set of keynotes didn’t draw nearly as many attendants as the first two, and reaction to the panel discussions was quite mixed. The first panel, which included senior executives from top software and hardware companies, pondered the idea of Java in 2005. Following the panelist’s predictions, moderator John Gage produced questions collected from the Web in the weeks prior to the conference. Panelists addressed concerns about the impact of Microsoft not supporting the 100% Pure Java effort and the potential for another technology to usurp Java’s position by 2005.

The insight provided gave one developer reason to cheer. “I like this type of talk because it gets me thinking about where I’m headed,” noted Michelle Stern, a Java developer from a Florida start-up. “I’ve invested a lot of myself (not to mention my time and money) into this technology, and it’s important that I see the support from the industry.”

This sentiment starkly contrasted comments I received about panel number two, which likened Java to an avalanche. The panel began with the Buttons and Threads Model demo by Miko Matsumura in which pairs of buttons were randomly connected by a thread. The panel, which was again moderated by John Gage and included Bill Joy, Matsumura, and author and researcher Stuart Kauffman, then launched into quite the esoteric discussion of the demo and its implications.

“Who is this guy [Kauffman]? I don’t understand a thing he’s saying!” “I feel like I’m eavesdropping on a private conversation.” “How does this topic apply to me? I thought keynotes were for general consumption.”

These comments were furnished outside the auditorium, after a mass exodus during the second panel made me curious enough to follow. When I asked resident mathematician and Java developer Daniel Steinberg about the Buttons and Threads Model demo that had me (and many others) so baffled, he chuckled and noted “I have a Ph.D. in math and couldn’t follow Miko’s explanation. The concept that he was trying to show can be explained in a much simpler manner.”

Likening the demo to the spread of a disease (which spreads as the people who have the disease spread it to those who don’t until there are a diminishing number of people left to infect and the spread slows way down), Steinberg explained that as each thread joins a pair of buttons, you begin to have small clusters of buttons that are joined by threads. Initially the average size of the clusters is small (four buttons or so), but as threads are added and clusters are joined growth really takes off. When the average cluster size reaches 20, it’s likely that the next few threads will join these big clusters to make even bigger clusters. Then, when the average size reaches about 200, the newly added threads will most likely join buttons already in the same cluster.

Still don’t get it? Check out the demo yourself! (See Resources below.)

Java, taking the world by storm

At least that’s how Geoffrey A. Moore, best selling author of Crossing the Chasm and Inside the Tornado (see Resources), might put it. The final keynote of this year’s JavaOne got a lot of folks excited by it’s driving concept: The Java platform is beginning to enter the phase (called the “tornado”) that will drive it into mainstream acceptance. In his talk, Moore discussed the evolution of the Java platform, in which he characterized the technology as a species that must survive in a world of competing technologies (what Darwin called survival of the fittest). He then illustrated the life cycle of technologies like Java, noting that once the technology has a gained the interest of key players it moves into a “chasm.” The tornado occurs once those players actually adopt the technology and drive it into the mainstream.

“If 14,000 developers, managers, and vendors aren’t evidence of mainstream acceptance, I’m not sure what it will take,” remarked a “computer science teacher, practitioner, and Java devotee” from Boise, Idaho. “We’re already a force to be reckoned with,” he continued, “and while Java may not deliver 100 percent on the ‘Write Once, Run Anywhere’ philosophy, it is the closest thing to it I have ever experienced. It usually takes me just a couple of minor adjustments to get my code running on multiple platforms. Compare this with [the] complete redesign of GUI interfaces, file interfaces, and network code required for other languages, and you see what I mean!”

Many other developers echoed these comments and condensed them into terms any manager would understand — bottom-line terms: “I am so much more productive in Java. I find myself putting off all other tasks, especially CGI, until all Java pieces are in place.”

Attendees offer insights and suggestions for next year’s show

Overall, the comments I heard were positive: “Not as much hype as last year,” “better choice of sessions,” “bigger venue,” and “quality speakers,” were just a few of the virtues of JavaOne 1998. But there’s always room for improvement, and several attendees offered some valid points about how things could be even better.

• Too many Sun employees as speakers. One vendor, who was originally slated (but later dropped from the schedule) to speak about the technology his company develops, felt that choosing experts from such a small pool diluted the effectiveness and coverage of topics, and excluded differing viewpoints on the technology.

• Too many sessions in too few days: “I’m glad they made the conference longer, but with so many tracks and sessions, I still found myself having to sacrifice certain sessions for others.” Is a week-long Javafest in order for 1999?

• Many sessions were entirely too crowded. A number of developers complained that seating was terribly limited in some sessions and BoFs: “You would think by now they would know which topics will attract the crowds.”

• Too much last-minute reorganizing. “I didn’t get here on time this morning, and now I have to scramble around to find out what has changed!”

JavaOne 1998 was terrific, but 1999 could top it. Are you listening Sun?