Morning keynote with IBM, featuring Erich Gamma and John Wiegand! Wow, haven't seen them since ECOOP 1992 in Geneva. Erich is a bit of a hero of mine, since he has done so many cool things over the years: ET++, patterns, Taligent, JUnit and Eclipse.
Surprisingly, what these guys did was describing and explaining the software release process at Eclipse! They did a really good job explaining why a six week delivery cycle may be perfect even for big projects like Eclipse. Are we witnessing the official death of the traditional waterfall model? I sure hope so. They finally introduced the Jazz project, which is a new set of Eclipse tools for supporting these processes. Erich then did a demo of some of these tools. I guess some people were bored to death, but I was pretty impressed. I'm just afraid the Rational part of IBM gets involved too much in this project. We don't want this to be a RUP tool. I didn't see any support for code reviews, similar to what Cenqua is doing with their Crucible tool.
All in all, impressive that IBM spent all their 45 minutes in front of everyone at JavaOne talking about processes and tools, not products.
Next up was Dynamically Typed Languages on the Java Platform by Gilad Bracha at Sun. He has the best title I've ever heard of; Computational Theologist. Anyway, he described the new instruction that will be added to the Dolphin version of the JVM; invokedynamic. I'm not overly interested in this area, but he also talked about hotswapping, i.e. changing code of an executing program, and he described briefly why this is so difficult for statically typed languages. An eyeopener for me.
I then went to a session with a pretty scary title, Java Technology, AJAX, Web 2.0 and SOA by a whole bunch of diverse people (Dion Hinchcliffe, Rod Johnson, John MacDonald, Craig McClanahan, and two guys from JackBe). I am, in general, skeptical to the whole idea about merging the server side SOA concept with the client side Ajax concept. They don't mix too well, I think, particularly since we have the general problem that you need a server side proxy to be able to call services (SOA) at a different domain from where a (Ajax) website originates.
Anyway, Rod Johnson started out great by saying that he was doing Ajax things in 1997. Sigh. In my eyes that knocked off several of his Spring karma points. It is totally irrelevant whether the Ajax technology has been around for a long time or not. Anyway, the whole session was a real yawn and the JackBe guys running it didn't do a good job. During the QA session there were lots of long-winding whacky questions that nobody understood. One of them sent Rod Johnson off on a long rambling why the dependency injection in Java EE 5 was inferior to Spring. Now, exactly in which way is this relevant to Ajax, Web 2.0 or SOA?
Next was a session describing new stuff in JDeveloper, you know that IDE from Oracle. I wasn't really interested, but all other sessions were even worse! Which is quite astonishing, considering... Less than 100 people attending, which is a good sign that their IDE will never take off. Sorry guys, your tool may be amazing, but peoples interests are elsewhere.