h1ghlevelb1ts

 

The first keynote was by an Oracle guy Mark Reinhold who sort of apologized for being with Oracle now. He went through the close future of Java and actually presented yet another schedule for version 7. This version that has been in the coming for so long will actually be split into to 2 and a smaller 7 will be shipped mid 2011. Some new stuff will be included which in my opinion only make the language look even worse. Generics and annotations were ugly moves but this new closure thing is even uglier. Better code in a language that supports it properly. Nice news on the other hand is the new property keyword to denote classic java bean properties. And the new stuff about modules looks good too. Kinda take some good parts from maven and put them in the standard. (The way Java has been driven all along - open source initiatives do good needed stuff which are complexified and put in the standard after a couple of years.)

 

I then went for a while to a rather good talk about the web but left halfway cause it was lacking new information for me.

 

Hardy has been talking about this guy that lives in Chania on Crete and makes a living out of a newsletter and consulting in the Java world. His name is Heinz Kabutz and he talked about reflection. He showed all kinds of crazy things that is possible to do with reflection - one scarier than the other.

 

Then a short sales talk by Matthew McCullough about git that was really nice. Didn't know about bisect and grep before so that was nice news. And that you can pull and push offline is pretty awesome too.

 

Sandwiches were hard to find half an hour into lunch but I finally got hold of an ok ham sandwich. The coffee was out though....

 

After lunch I decided to find out what the next big JVM language is. Stephen Colebourne made a great job of listing a bunch of essential criteria to look for in the next language and then went on to disqualify clojure, groovy, scala and fantom. He was a bit for the 2 latter ones mostly because they are statically typed. Statically typed rules because it forces compilation upon lousy developers of which there are many. I tend to agree with this. Developers without unit testing abilities should use statically typed languages. Well then - if neither of these 4 languages don't cut it.... a new version of Java without backwards compatibility could be the answer. Stephen went on to argue that if ORACLE could increase the productivity of their 10000 developers with 5 % they would indeed make even bigger piles of money.

 

Then a rather technical talk about Cassandra by Jonathan Ellis. A bit over my head really. It is a bit symptomatic that NoSQL talks typically is about how the database has been optimized for scalability. This is probably because there is nothing else to say...

 

Scala after that. First a fun session with Dick Wall and Bill Venners about what is new in 2.8 and then a session on parallellism by Odersky himself. Both good. It is a bit interesting to notice that all Scala presentation I have been too this far (5 I think) have all put emphasize on scala advocacy regardless of their subject. Look how bad it is in Java and how sweet it is in Scala kinda. There are lots of really cool stuff in Scala and it is getting better for every new version. What I don't like is that there are both good and bad ways to do stuff. But I definitly have to learn it a bit more before judging.

 

Then beers! Had a really fun evening with Rickard, Sofia and Petter at Kulminator a bit south of the city centre. La Trappe Quercus was probably the best beer I have ever had.

 

Thursday morning was not the most alert I had had so I slept a bit longer and then made some final preparations for my own talk before walking another way through the docklands to the venue. So I missed the J2EE keynote - not particularly sorry about that. The first session I went to was right before my own and was about OpenJDK. Dalibor Topic made a good job of describing the work being done and what it would take to be a contributor. Would kinda cool wouldn't it.

 

My own talk on continuous integration went ok. It is always a bit awkward to speak in english for me but I did okey most of the time. I think my message was clear and I also managed to stay within the 15 minutes without problem. Will I submit talks to next year as well? Maybe. Probably about something more technical.

 

After that I was really tired and dozed through Monster Builds and How to Tame Them by Chris Mountford from Atlassian (sponsor of beer!). He sort of continued with what I had talked about and delved into it more deeply. It was a good and inspiring talk.

 

I didn't expect much from my first encounter with the Java Posse so I was not disappointed. It was a fun session with key bits of information delivered in a collaborative and entertaining way.

 

Todays top speak was surely Neal Fords talk about emerging design. .....

 

The day was rounded off with a bunch of Java puzzles delivered by Joshua Bloch and William Pugh. This was a number of Java code bits with strange behavior. It was a fun session that unfortunatly missed the punch line. Lessons learned was strangely enough not about how unit testing might have detected these strangenesses. Nevertheless fun and interesting. The mess of the implementation of generics was clearly revealed once again.

 

The final day started with a round table discussion with several java dignitaries. A mostly fun and somewhat interesting session although I dozed off for a while. When writing the last words of this post I am listening to a session about Apache Camel. After this it is time for some serious sightseeing.

 

The venue is really suitable for this kind of conference. It gets crowded when 3000 nerds switch rooms at the same time but it has been working fine. The food has been ordinary but that's to be expected at this big a venue. Antwerpen is a fascinating city with a brutal blend of old and new. And of course the beer is really something.