JavaOne 2011 got off to a bit of a shaky start this morning with there being a lack of seating in the Grand Ballroom of the Hilton, leading to the escalators eventually being blocked by physically meek, yet surly, security guards and having people being redirected to smaller rooms somewhere else in the rabbit warren that is the Hilton’s conference and event space. However that didn’t happen until after a couple hundred poor souls were left standing at the back of the room to endure a 2 hour-long dry and technical keynote.
Then Mark Reinhold went missing. Not sure what happened there but awkwardly is name was announced and someone else took the stage with no real explanation about why. He then went on to introduce Doug Fisher, VP Intel, who was supposed to be the 2nd half of the keynote. The Intel guys and their Oracle counterparts presented myriad of numbers and graphs to prove that Java runs well on the Intel architecture. Not really sure anyone needed a lot of convincing of that, but their results were impressive nonetheless.
Eventually Reinhold did appear to give an overview of the JavaSE 7 release that occurred during 2011, with its focus on Project Coin. Followed by a quick nod to the Charles Nutter and the JRuby guys, who apparently thought they were coming to a college frat party and dressed appropriately. Apparently the distinction between cool and unprofessional has not yet been fully grasped in the JRuby offices.
Reinhold then moved on to features coming in JavaSE 8 which is scheduled for release in 2012. At the top of the list is Project Lambda (JSR-335) (ie. Closures). This is followed by Project Jigsaw, which is the JavaSE team’s attempt to catch up with the OSGi community and offer better modularization of applications and as a result, clean up the mess that is the Java classpath.
Next there was some sweeping ideas about what might be in Java 9, but I think based on the debacle that was the delayed release of JavaSE 7, the Java development community is a little skeptical of such long-term predictions.
Next Linda Dimichel came to the stage to talk about JavaEE 7. She went straight for the big buzzword of the conference, the cloud. Some very interesting ideas around building in PaaS support directly into JavaEE. The good news is that it sounds like Java PaaS has been validated by Oracle as a genuine industry direction. The bad news is that while adding PaaS features and support at the JavaEE layer might help eliminate vendor lock in and other related concerns, it has even more chance of just creating a horrible, inefficient mess of standards and code that vendors will support to varying degrees and applications will not become any more portable in the end. Dimichel did commit to a transparent process for defining JavaEE 7, so hopefully this will help to steer the release away from some of the bigger land mines.
The big question about the JavaEE 7 PaaS and Cloud focus is what will it mean for the existing vendors in the space (CloudBees, Amazon, VMWare etc)? Are companies going to roll their own PaaS based on vanilla JavaEE 7, or they still going to rely on 3rd party vendors. I think the latter is more likely, as running a production ready PaaS environment is a pretty specialized skill set that most companies will not want to take on the expense of. The biggest problem for the vendors might be the unrealistic expectations set by the Oracle marketing juggernaut about what PaaS can be and what JavaEE 7 can let them do.
Finally we moved on to Java Mobile and Embedded with Hinkmond Wong. I suspect if it wasn’t for the ongoing lawsuit between Oracle and Google, this part of the presentation would have been just a single slide that stated “Please just give up and use Android so we can move these resources on to other projects, Thanks Oracle”.This was mercifully the end of the keynote that went almost 2 and a half hours in total.