I have been to FOSDEM 2005 this weekend. fukami has already blogged about it. It’s a meeting for open source developers and quite a lot showed up (although I doubt the official number of 3500, it was closer to 1500 to 2000 for me – but that’s just my impression). So the event is about development and so is this

The event is free (as in beer) for the participants although you are kindly asked to leave a donation. They are selling expensive t-shirts to give you options. The venue is one of the ugly buildings of the Université libre de Bruxelles (ULB). Quite a lot of rooms for people to talk, not much space for exhibition which is a pity. Also the exhibition area could have been a bit warmer, but I doubt there was a chance to do anything about that as it was fucking cold this weekend.

One of the bigger annoyances was Richard Stallman who talked about possible changes to copyright law. While his talk was not totally disappointing it didn’t deliver any glorious insights worth mentioning. However, I could have lived with his statements if he would have been open to comments and discussion. But when people started asking questions he exposed an unbelievable arrogance. When people haven’t been clear enough (mainly general english problems – not unlikely to happen for non-native speakers) he refused to answer because he „can’t understand the person“ talking to them third person. Such an asshole…

Jimbo Wales‚ introductory talk about Wikipedia was fine. He adapted well to the audience and pushed the projects main problems they face at the time: developers. They do not have enough developers to deal with problems like the overall growth in terms of users and data, the overwhelming traffic and the need for more sophisticated features in the software, especially to deal with meta-data.

Alan Cox‘ talk on his personal approach to Linux kernel development was a revelation. For a while now, he produces his own version of the Linux kernel dubbed -ac. this kernel is an alternative to the „official“ kernels of Linus Torvalds that usually introduce a ton of changes that tend to break a lot of things. In order to compensate, the -ac kernel tries to be more conservative. He doesn’t use the same revision system, he applies patches depending on taste and his tools are diff, patch and evolution („the search engine works for me“). Impressive, but the Linux kernel development model is a mess.

I hope, the public will wake up one time and shift the hype to the threeBSDsystems, where it belongs. Their development model is by far more professional as is the code that results from this model.