log4p

Wicket 2 (Martijn Dashorst)

I knew that a talk labeled ‘Wicket 2’ was a bit suspicious. There is no such thing (anymore) as Wicket 2, the latest release is 1.3 and upcoming is 1.4. I have played around with Wicket a bit in the past and knew a bit about it at forehand. Martijn didn’t really expand my knowledge. He did however do some live coding, which in concept is a very good idea. In practice it wasn’t such a great one. Martijn managed to make typing mistakes in almost every line he wrote which really slowed down the presentation.

Still, I think Wicket is a great framework; I like the error screens (which we saw to many during the presentation) in particular which quite accurately manage to tell you what is wrong.

Terracotta (Geert Bevin)

I’ve seen presentations on Terracotta before and I’ve seen Geert Bevin presenting (Rife) before. Both where great. Somehow the combination didn’t really add extra value. The examples where impressive again. When will we see the first application servers with Terracotta build-in?

OSGI (Peter Kriens)

Peter Kriens managed to make terrible jokes and give a far to complex presentation about the excellent OSGi framework in the same time. I really like the concepts in OSGi and think it will be everywhere in the future.

Test driven development in practice (Lasse Koskela)

Lasse gave a nice and solid presentation on TDD. He started with the TDD mantra:

After the introductio he talked about one of my favorite concepts: behavior driven design. He actually didn’t understand how it differed from TDD. And conceptually I t

hink he is right. As long as you test behavior using unittests there aren’t much/any differences. Personally I find it easier to think in terms of behavior when using a DSL to do so.

He also talked about test doubles, and what type of test doubles should be used in which situation. The types he talked about where:

1. Stubs (simplest possible thing that could possible work)
2. Fakes (alternative/lightweight implementation)
3. Mocks (interaction based)

His rule of thumb would be to start of with a stub, use fakes if stubs are not sufficient anymore and start using mocks if the other two approaches are to limiting.

Now, finally, during the last session they actually started programming on stage! It was a bit slow… but it was a very nice way to end the final session of the conference.

Special thanks

Special thanks go to Marcel Maatkamp for never leaving home unprepared. I probably wouldn’t have found my car-key after dropping it if he hadn’t been carrying a Maglite around!