Day 2 started off with some coffee and a session titled “SOA using Java Web Services”. I must say that I had high expectations for this talk as I recently had been in lots and lots of discussions about SOA and web services in general. This talk – at least by the title and the abstract – seemed to be interesting in that light. Besides, the talk was done by Mark Hansen who also authored a book with the same title and I knew some of my colleagues really appreciated that book. The session itself turned out to be a deception. Mark was clearly not prepared for the talk. His demos were not setup and failed when he tried to get them to run anyway. The title and the topics on the agenda seemed interesting at first. Topics like Web Service Platform Architecture and parts on REST and JAX-WS looked promising, but he managed to translate such topics into very basic 101-talk. Apparently he assumed the people in the audience to have never heard about web service at all. Then, only minutes later he assumed we were all subject matter experts as he assumed that everybody in the audience by now definitely knew the difference between REST and SOAP. Not that I didn’t, but the swings in the level of the talk were strange to say the least.
Actually the only thing I took from the talk was a one-liner: “sometimes it seems as if the ease-of-use effort serves the single purpose of ease-of-demoing”. I think he’s very right with that. During the break I decided to switch to the session on Flex with Bruce Eckel and James Ward. Much, much better presenters and a nice talk that I could seamlessly blend in although I’d missed the first part. They pretty much kept it at basic stuff and some demos. No real advanced or mouthwatering GUI stuff that I expected to see. But then again there are more Flex sessions this week, so I think the fair share of GUI extravaganza is still to come. The big difference with the web services talk was that this talk made sense.
The afternoon session I picked was by Kenneth Saks and Linda DeMichiel and covered practical programming with EJB 3 and JPA. Actually this session was some sort of a deep dive into the darker parts of the specification. Ken started off with Session Bean related topics, like e.g. distinction between the component environment and the global JNDI namespace, intra-component method delegation, transactional behavior of Timers, application clients and mixing EJB 2.x and 3.0 technologies. After that Linda took the stage and talked amongst others about the removal of orphan entities, inheritance hierarchies, and fetching strategies. Especially the latter made me think. The examples illustrated the impact of the fetching strategy on the number of (unnecessary) joins and database roundtrips. She tested the audience by asking some trick questions about it. Great stuff!
The university part of JavaPolis ended with two shorter sessions just like the day before. Only this time I took a two part session on Jazz by Erich Gamma. I’d seen Erich speak about Jazz the previous year and a few weeks back I attended a session at IBM where Jazz was also one of the topics. Erich didn’t really show any new stuff that I hadn’t already seen but did a nice talk on the rationale behind the whole idea and how it’s coming down. The second part of the talk was supposed to be a live demo but because of some network router problems this turned into Erich showing some videos from his laptop with him doing live voice-overs. The session once again fueled my mixed feelings on the subject. On the one hand it all seems like a great idea. Team-first development and a set of integrated tools that support the process of choice instead of dictating it. On the other hand IBM is determined to ship it as a commercial product. And worse, the real power of Jazz only comes with the total stack that offers maximum integration. So if you’re stuck on another version control system or another issue tracker you get much less out of it.
The story sounds a lot like they are trying to copy the Eclipse success but since this is a commercial product I’m really wondering if the market acceptance is going to be huge. All in all I have to admit that on slideware it is a very interesting product so I’m definitely going to sort it out.
With that another day of JavaPolis University ended, but the highlight of the day had yet to come. As I wrote yesterday, I was invited to the JavaPolis Speakers and JUG Leaders diner, being a co-lead of NLJUG myself. As I walked towards the restaurant a taxi pulled over and James Gosling himself stepped out of it. Not everyday you meet people like this. But James was not the only one present. About 100 speakers and JUG leaders joined the party and there were lots of people that you normally don’t have access to.
Discussions were diverse e.g. OpenJDK, JRuby, fragmentation in the mobile world, the iPhone, and of course enterprise computing in every aspect. As the restaurant turned out to be a former Beer Brewery you can image there was lots of it. Unfortunately I came by car so I didn’t want to risk a crash somewhere in the center of Antwerp. Other people didn’t have that problem so conversations got lively 😉 A lot of funky pictures were taken at the party and even some video. So right now I’m wondering where this stuff will be showing up 🙂
Lastly, two more things to say on my previous blog post:
When I read the posted entry on the web I noticed some *** in the text. It got me puzzled for a second, but then I realized it was the dirty words filter that we installed. I promised myself not to joke about it, but a name like D*i*c*k Wall does not seem to pass the filter 🙂
And the other thing…I got the link to Paul’s blog wrong. I sort of deducted it from the URL to my own blog, but it appears that Paul got a totally different URL. Don’t know why but it’s fixed now.