JavaZone is over. I don’t have many other conferences to compare it with, but I thought it was excellent. The following is a short summary of what I attended. I have a lousy memory so I should have written stuff down, but I’m too lazy. 😉
- Test-Driven Development – Johannes Brodwall and Trond Pedersen – Basic but great introduction to TDD. Nothing fancy. I thought it was good as I am a total rookie on TDD, and they cleared up some important aspects of it. Writing tests to describe bugs and missing features is a mindset I’m not used to. I’d like to try TDD somewhere in the near future.
- The secrets of IoC – The PicoContainer advantage – Aslak Hellesøy – I was a bit disappointed about this really. I pretty much knew what IoC was and how it works. What I wanted was to know why it is so great, and what problems it would solve. Aslak explained the basic concepts and how Pico- and Nano-container worked. The selling point was that it was great because it is now testable. I know that’s a big deal for Aslak because he’s an avid TDD fan. But it’s not really the best selling point for me. What convinced me about a month back was this blog by Rickard Öberg (which also attended JavaZone).
- Model Driven Development: Who Needs It – Wim Bast – Wim was probably the only MDA guy at the whole conference, and he gave a good introduction to MDA and its concepts. I’ve been a bit sceptical, and still i still am, but his introduction made me understand the basics. The rather naive image I had of MDA earlier where I imagined they drew a lot of technology independent diagrams and then generated the code was changed. Of course it could resemble this, but it actually has some steps of refinement, two of them. The first one is a refinement of the domain model considering the technology, and the second refinement is the transformation into code. These refinement levels as well as the domain model will be tweaked until a satisfactory result has been reached. Of course not all code will be generated, but this led me to understand that MDA might work some day, at least for some applications. I’m not sure I will be a user.
- Open Source Agile Architecture – Trond Arve Wasskog – This was the best talk I attended. He went through a concrete project that has used WebWork2, Spring and Hibernate. It was absolutely brilliant, I want to go out and base my next project on these technologies. 😉 He talked about how WebWork ensured testability for your actions, as well as Spring and IoC did for your business logic. But what really catched my interest was the transaction handling and security possibilities that AOP brings to Spring. I don’t know much EJB, but the only arguments I usually hear for them is transactions and security. I’m going to insist on Spring next time around. 😉
- A J2EE middleware migration – Kristian Schjelderup and Otto Fjøsne – Telenor has migrated it’s middleware from WebSphere 3.x to 5.x. In an big organisation like this that is a hughe undertaking. They had probably around 40 backend systems (this is where I should have taken notes 😉 ) and even more frontend systems, and needed to migrate. During that migration they also decided to restructure everything. This meant breaking package names, creating and restructuring modules and creating a better middleware that frontend applications could use to access the backend systems. They put down a core team that was responsible for the guidelines and doing QA to enhance the quality of the migrated platform. This was really a complex undertaking and the key to it all was the core team, simplicity and good guidelines. The migration was done gradually and no system on the old platform was shut down before the new system ran correctly on the new. It was interesting to see how they handled it all, but I’m not sure most organisations would realise the benefits of such an project.
- The Java Trap – Richard Stallman – What to say? Stallman is Stallman I guess. All software should be free. It was entertaining, and it would be a nice world if all software was free. I’m not sure I agree. 🙂 He’s focused a lot on the analogy of “helping your neighbour”, and that’s good really. But what annoys me is that it is mainly based on feelings. We all want to do good, but good can be subjective.
Allright. That’s it. I really enjoyed myself. Some beers at night and a nice but unstructured BOF on the future of development. I’ll be back next year.