This general session concentrated major news about JDK (desktop?), Java EE, and Java for Mobile and Embedded. Mark Reinhold basically explained the roadmap for JDK releases, which will follow Plan B. This means that there will be a release in 2011 with the features that are ready, and another release in 2012 with the features that are still work in progress at that stage. I will not into details, refer to Mark Reinhold’s blog for a list of features of the upcoming JDK 7.
I will only say that the functional programming enthusiasts will have to wait, as Project Lambda, which plans to introduce closures into Java, has been deferred to JDK 8.
Throughout his presentation, Mark Reinhold has been very careful to show that all the major research and development work is performed in an open manner, and that the JCP procedures will be followed before to include anything in the language. Oracle is trying to prove that it supports the language.
About Java EE, no big news, as version 6 is less than one year old, and products are still being developed by the platform vendors. Of course, like in all other talks, there has been a reference to a Web tier, with a Web profile, RESTful services, HTML5 and JSON support and other usual suspects. Talk about strong directions … At least, the message is consistent.
Newt came Greg Bollella, talking about mobile and embedded stuff. He basically started with a statement from Oracle:
Oracle is committed to modernizing the Java ME platform
OK, message received. In practice, we are looking at an upcoming Java ME.Next, which is not yet clearly specified, expect for a few principles:
- Updates across the board (Virtual Machine, API, framework)
- Guaranteed backward compatibility
- Optional packages remain present for flexibility
- JCP principles will be followed
The Web extensions for mobile applications that I outlined earlier is also in the program. However, I got disappointed, as no formal roadmap was announced. Since they are at the prototype level right now, I am afraid that it is still going to take years to iron out the “details” and make it a final specification. Too bad, we would need that quite soon.
Then, Greg Bollella moved to his favorite topis: embedded systems. He defined embedded as covering a wide area of things that contain a processor, with software, and that are not considered as computers. With this definition, there are 1000 times more processors sold in that space than in the the workstation space. Yep, that’s a big market.
According to Greg Bollella, another trend is that all these embedded devices are producing operational data that is just not used, because there is no way to get this data anywhere. For instance, in a car, many parameters are recorded every few milliseconds, temporarily stored, and then discarded. Most likely, it could be used for something.
After building all the excitement, there was nothing real to announce. Of course, the new Java ME is going to cover these embedded systems, but we didn’t get anything resembling a feature set or a roadmap. Things are hard on the device space.
Bollella mentioned Project Verrazano, though. It is a tool that takes a JAR and a platform spec, and it generates a self-contained output JAR, with everything required, including the APIs, the VM, and whatever else is required. We learnt that HelloWorld on Java 6 can be shrunk from 58Mb to 8Mb, a 80% reduction, but still rather big (apparently, this is due at least in part to the fact that native code is not filtered intelligently). Of course, this is intended as an analysis tool, in order to determine which parts of Java are used (or not used) in embedded application, and to derive from that a specific profile. I guess that an adaptation of this project to the Java Card 3.0 Classic platform would be very practical for the generation of static Java Card products. However, I haven’t found links to this project on the Web. Let’s wait and see what happens.
Overall, not a bad general session. Oracle made great efforts to display support for Java, but the announcements were not really groundbreaking. One positive note, though: it gives hope that there will still be a JavaOne or similar event next year.