Posted by sonyabarry
on March 20, 2011 at 8:56 PM PDT
Last week I attended The Server Side Java Symposium in Las Vegas. It was a great conference, with a nice mix of technical sessions and larger discussions of the language, the community, and the JCP.
Wednesday morning started with a little excitement - a site-wide power outage at the hotel. It's a little weird when a giant hotel is completely dark and completely silent. Fortunately they got the power back on about 20 minutes before they opened conference registration that morning.
This was my first time attending this conference, and I have to say I was impressed, it's small but everybody here seems very engaged and the sessions were interesting - a good mix of technical content and discussions of larger issues and the future of Java.
James Gosling did the opening keynote Wednesday morning. I wasn't sure what to expect. He always gives a good talk, and it's been a while since I've seen him at an event. He didn't have a whole lot to say about Java technically at this point, which is not surprising given that he's not been involved with Java at all for the last year. I'm sure he hears things through the grapevine faster than a lot of us, but he's on the sidelines now too.
One thing he said that I think rings true is that at this point the JVM is more important than Java, because it allows us to use the right language for the task at hand. That's a position that I've taken for a while now, and I realize that it's a little controversial. I don't think it's an insult to Java, but a support to the ecosystem in which Java lives, and a support to Java as well. Java can't be all things to all people - it will be much stronger in the long run if development continues to build on the things that it does well, and room is left for other languages to fill in the gaps that Java just can't cover. He also talked a lot about different flavors of clouds, and the privacy and regulatory issues inherent in developing in and for them.
Steve Harris (Sr. VP App Server Development, Oracle) and Adam Messinger (VP of Development in the Fusion Middleware Group, Oracle) did a keynote called Java in Flux. They went over a brief history of the last year of Java at Oracle and laid out some plans for the future, including highlights of JDK7 (COIN, dynamic language support, and concurrency and collections updates, among others), the direction they're moving for JDK 8, and even a wishlist for JDK9. They also covered the current and immediate future of JavaFX and Mobile, and talked a bit about Java in the cloud as well.
We closed the evening with an Oracle sponsored Ask the Executives panel, including Steve Harris, Adam Messinger, Patrick Curran (Chair of the JCP), John Jullion-Ceccarelli (Senior Engineering Manager of NetBeans and a few other things), and Justin Kestelyn, (Oracle Technology Network, and my boss).
One person asked what percentage of original Sun engineers are still on board at Oracle. In JavaSE Adam said that attrition has been low - despite the notable departures, the team is still about 90% people who came from Sun at the time of the acquisition. John made the point that there was a lot of attrition between the time the acquisition was announced and closed, but it's not unusual for a big event like this being an impetus for people who have been complacent in a career on one project for several years to move on. I'll add that I think that's true, and there were also a lot of layoffs in those last years that were done by Sun, not by Oracle.
A question came up about vendor lock-in of Java in the next five years. There is a concern about less transparency and less openness in the future. Patrick Curran addressed it by talking about the well defined process of the JCP, and pointed out that the fact that the JCP has been stuck for a while indicates the exact opposite, since other participants in the JCP were able to block things that both Sun and Oracle have wanted to do.
Thursday morning we kicked off with a panel discussion of the JCP. The panel was Reza Rahman, Patrick Curran, and James Gosling. Gosling talked a bit about the history of JCP. He said that Sun had always worked in collaboration with industry and universities, and was involved in various consortiums. They were aware of the good points and bad points, and wanted to run Java the same way. At some point a selection of partners were suspicious that Sun wasn't acting fairly, so they requested some formalization and JCPP 1.0 was born. At that time the expert groups were entirely engineers. They desired extreme transparency and tried to audit all decisions. That didn't work so well in practice, so with JCP 2, auditing disappeared and the executive council was born.
So when did it go off the rails? Reza described process driven issues, management issues, and some issues that are just "the nature of the beast." The next questions was what is different between the JCP at Oracle and Sun. According to Patrick Curran the main difference is that Sun got stuck and wasn't willing to make the compromises or decisions necessary to move forward. Oracle committed to getting unstuck and things are moving again. Policy-wise nothing has really changed - Oracle is being careful to respect the licensing and process that Sun was using.
Gosling agreed that much is the same since the transition. Reza observed that Oracle engineers seem to be a little more practical than some of the Sun folks which can help move things along in the real world. Oracle spec leads run 2/3 of the JSRs. Historically that's been true. Are there plans to hand some of that off? Yes, they said that they would like to spread that out, but it requires a lot of time and energy commitments. Regardless of who the spec lead is, the actual design work is done by the expert group, which is not entirely Oracle.
Someone asked what small things can people do who want to be involved but don't have a ton of time. The answer was simple:
1. follow the thing that you're interested in
2. read the spec and comment on it (you can do that without joining)
3. play with the code if you have time
According to Patrick Curran most spec leads are desperate for feedback, and Gosling commented that a spec will come out and people will complain, but nobody brought anything up during the review cycles, before it was too late.
During the conference I also caught some technical sessions with Steve Chin, Kirk Pepperdine, and Adam Bien. It was all too much for me to cover here, but you can take a look at the slides and videos at the links below to see what was going on:
Slidedecks from presentations are here: http://www.slideshare.net/javasymposium
TheServerSide's own coverage starts here: http://www.theserverside.com/news/2240033365/TheServerSide-Java-Symposiu...
I thought they were taking audio recordings of the keynotes and panel discussions, which I hope they will post links to from their coverage page.