0 comments
Java 
Well, here I am, for my first time in San Francisco, at my first JavaOne! Great weather, nice city, good local beer (Anchor Steam). So much to do and see, so I barely even have time to blog. So today two days at once: let's start with monday, the university track (day 0).
The university day is a pre-conference day filled with longer tracks to provide more in-depth information on certain topics.
The university got started with a key note by Rich Green and James Gosling. Gosling had an interesting remark on how software vendors like Sun keep promising to make software development easier, but in practice, it doesn't. Basically he compares it to the Peter Principle: software development just keeps getting pushed to its limits. When a shiny new thing makes existing problems easier to solve, the customer/boss keeps adding new todo items to the list of the software developer, so in the end, he/she still has the same amount of work. It made sense, though somehow I think not only the customer/boss, but also the software vendors push developers to the next problem, otherwise, who will need the next version of their products?
The morning session I attended was "Filty Rich Clients" by Chet Haasse and Romain Guy. Basically Swing on steroids, Chet and Romain gave some good insights on how to do some cool animations and effects, while still keeping an eye on performance and maintainability. Individually, they are good speakers, but as a team, they are even better. They both written a book on the subject which has the same name as the presentation and will come out soon. I am definitely interested, the session showed that Chet and Romain really know what people are doing with Swing, and what the common pitfalls are.
I knew in advance the afternoon session it was going to be tougher to find something interesting. I tried SOA patterns for about 5 minutes, but I just went crazy. The speaker was just yelling like a madman, trying to reach a speed of explaining 10 patterns a second. I switched to a track on Web Services and this was a lot better. Especially the guys on the WSIT team gave some good insights on where Web Services interop is going. Finally, interop between Java and .NET is taken seriously on both ends, and it shows. The more advanced WS-* stuff, like WS-Policy, WS-Trust and WS-Coordination are working pretty much seamless.
In the evening session I attended a talk on DTrace. It was a crazy talk given by two British Solaris kernel engineers. Their laptops where still on GMT time zone, and so where they, judging on the amout of Red Bull on the table and all the crazy (live) demo's. The talk was really interesting and fun, DTrace has moved along well the last couple of years. There is a Java API now to use as a DTrace consumer and a tool is provided called Chime. Chime is a GUI application that allows some basic DTrace monitoring without having to use scripting or console magic. It's pretty useful to get started with DTrace. Being kernel engineers though, the Britisch duo didn't like GUIs, and rather show off some assembler to show how DTrace did its magic (after sitting through 10 hours of presentations, that was only spent on geeks like myself). Being such a fantastic tool, it's great to see that DTrace will be part of the next MacOS update, and is also part of FreeBSD (and of course Solaris 10/11).
All in all a very entertaining pre-conference day. The only complaints I have are lack of caffeine (only decaf available in the afternoon!), and not being able to enjoy the nice weather that much. On to the next day!