LarsWestergren's Journal: Java One 2007 wrapup

I had intended to write a daily report on Java One like Cay Horstmann or my colleague Ola Bini , but I was too exhausted to do that. Now we are driving up to Portland for RailsConf, and I finally have the time and energy to write something.

This was my second Java One. This one was better than last years in my opinion. Either the quality was higher or I've become better at picking the interesting sessions .Seeing the amount of innovation going on, I'm not worried that Java the language or especially the platform will disappear anytime within the next 10 years (if it remains the dominant language remains to be seen of course). Seems there is lots and lots of cool stuff going on in robotics, embedded and mobile markets. Not something I work with, but the technical demos last days by Perrone Robotics, ABB and so on were really impressive.

This year I went to several labs. DTrace lab was great, I'm installing OpenSolaris or Nexenta as soon as I get home and will try to contribute to JRuby and other open source projects with DTrace. SunSPOT lab was also great fun. I'm buying a dev kit as soon as they are legally available to order in Europe. The currently manufactured batch did not comply with EU laws for decreasing dangerous chemicals in manufacturing process. I like those laws, so I'll take the minor inconvenience of waiting a month or two for my new toy rather than circumvent import rules.

JavaFX (previously F3 or Form Follows Function) was neat, but I see it is facing a uphill battle. Some say it is not meant to compete with Flash or Microsoft Silverlight, but it will obviously be compared with them anyway and it seems tool support is currently lacking for JavaFX to be able to attract large numbers of graphic designers or developers. However, perhaps it will at least prevent Java developers from leaving the platform. Some have also questioned making a new language for it and have suggested a better solution would have been to use Groovy, JRuby or any of the other existing scripting languages to create a Domain Specific Language instead. Perhaps, I'm not familiar enough with dynamic languages to know if that is a superior solution. Also, remember that JavaFX have been in development for almost three years (I believe), and when they got started the speed and stability of dynamic languages on the JVM was nowhere near where it is today.

I went to several technical sessoins about new features in Java. The Superpackages, modularity and JVM Kernel technical sessions were interesting. If Sun and the open source contributors manage to pull it off quickly and correctly, Java the platform will really be a force to be reckoned with on the desktop. Neal Gafters Closures session (my last TS of the conference) was really good. I really really want closures for Java now, and I think the BGGA proposal comes closest to what I want, good trade off between backward compatibility and power. I don't know if everyone in the audience understood that with closures, we wouldn't have to write this kind of stuff ever again in Java:

try {
    inStream.readSomething();
} finally {
    try {
        inStream.close();
    } catch (IOException ex) { //Comment to stop Checkstyle or PMD from screaming at me for having an empty catchblock...
    }
}

Instead you would import a util class somewhere from java.io, that class defines a "withStream" method that takes a number of argments AND a closure argument, and that util class takes care of all the boilerplate code for you once and for all, and using it in your code would then just be to use the method something like this instead (sorry if I get the exact syntax wrong):

withStream(inStream) {
    inStream.readSomething();
}

You can similarly use it to simplify concurrency, iterations, or anything that you would previously have used anonymous classes for before, but closures don't have many of the problems that anonymous classes do. With closures, you can refer to variables outside the closure without them having been declared final, you can use "this" and "return" keywords just like normally, you can nest closures and it still readable, and so on. Now that we are on the topic of that technical session, it seems that feelings were a lot more heated at this years Java One. I didn't stay through the whole Q&A, but from what I have heard from others the discussion at this and other sessions got really heated. Not to the level of screaming and swearing, but definitely tense. It is good that people are passionate, but I hope it doesn't turn into personal attacks.

I think there was less "swag" this year, but that's ok by me. In my opinion much of it is junk that goes straight to landfill after the conference anyway, and chasing after prizes distracts from what is important at the conference - making new contacts, participating in the open community, and learning stuff. That said, the high end prizes were probably better this year. $5000 for the virtual flying dukes contest for instance. I didn't submit my entry, my avatar only managed to catch the t-shirt 50% of the time, and I thought it had to be 100%. I went to the jMonkeyengine session on Friday, and it turned out I had made the maths much more complicated than needed. A bit depressing that I didn't manage to solve it considering how much math I've been doing last year. Oh well, better luck next year. The avatars people had been doing were not all that impressive. In my own humble opinion, the stuff I had done in Blender was much cooler. I will continue practicing, and we shall see next year (if I get to go to JavaOne again that is). They said the competition would probably return in some way,

Most important thing at the event for me was probably the open sourcing of the java platform. The people at Sun are really dedicated to this, and things are moving ahead quickly. I met people from JMX in person, which was great, and talked to lots of other interesting people. And I won an iPod thanks to being the fastest at solving an anagram in a room of around 75-100 people at the Sun Certified Professionals party. Woo!