Ubuntu Dev Summit in Paris.
So this release cycle's Ubuntu Dev Summit has ended. 60 people from the wider Ubuntu community gathered in Paris to draft the next release, Edgy Eft -- and it will be rocking (although probably edgy, as its name says). Some highlights are an event-based replacement for the init system, SysV init does not fit today's use cases very well anymore, a system nowadays is almost never booted with all kinds of pluggable hardware, suspend and resume functionality. Speaking of suspend, I've talked to Ben Collins, Ubuntu's kernel maintainer about integrating suspend2 in Edgy's kernel. He had a quick look at the patches and was quite surprised by the improved quality. Nigel Cunningham, suspend2's main developer has put quite some work in making the patches less intrusive and more mergeable. Last week, I benchmarked both, suspend2 and vanilla's software suspend. The results were simply amazing: swsusp needs 45 seconds to bring the machine back up, whereas suspend2 does that in 25s (mainly caused by image compression). suspend2 also adds a nice userinterface (framebuffer splash, textmode progress bar -- pick your poisoln) and is much more reliable under high memory load. The least thing you want is a system not suspending when you run away in a hurry and drop your laptop in a train. That's definitely a big plus. Hope Ubuntu will be the first distro shipping this cutting edge technology by default.
We also talked about the benefits and disadvantages of decoupling Ubuntu's and Kubuntu's release schedules. It would make sense from a branding point of view. Putting out a press release of Ubuntu and Kubuntu at the same time doesn't make sense, it'd only harm the clarity of the message to the outside. That is mainly why Kubuntu didn't have a separate press release. To create a stronger brand, however, it'd make sense to have a real opportunity to notify the public. From a technical point of view, releasing for example 3 weeks after a Ubuntu release makes it possible to get the first wave of bugfixes onto the CD, so one might expect to be of higher quality. I talked about that with Jonathan of Kubuntu fame, and also with a couple of guys that'd probably be affected by it, notably Daniel Silverstone from the Soyuz team (package archives infrastructure), and Colin Watson and Matt Zimmerman from the release team. They were not directly opposed and certainly see that it'd add benefit. The details however need to be sorted out first.
One thing I'm really excited about is the level of cross-desktop collaboration I was confronted with last week. On Thursday, we had a BoF that was called gst-umbrella. We talked about how GNOME system tools could be improved, and how we can share more code between the three desktops living in the Ubuntu world -- KDE, GNOME and XFCE. The atmosphere was really relaxed and we were very much on the same line. There is quite an opportunity to use the same configuration manipulation backends for all those desktops. Currently, according to the GNOME guys, the GNOME system tools source package is a tarball of 12MB of perl backends, C frontends and XML-based GUI files. Feel the pain. There is a common understanding that it makes sense to move that stuff to Python in the future. Those issues have to be checked back with GNOME-upstream. We might even have the first results of that ready for edgy, which has a pretty tight release schedule -- feature freeze in 11 weeks.
The people that might worry about me after Ken's
report from wednesday night in Paris : I'm fine. Only have to wear my (prescribed)
sunglasses now all day (and night!) which makes me look a bit stupid. Can't say I'm
unhappy with that acutally, people should basically never expect me to be too serious. :P Still a
good thing that it's really summer here in Western Europe.
[ Sat, 24 Jun 2006 12:29:32 +0200 ] permanent link
This weblog does currently not offer the option to comment. I would be happy to receive an email with your thoughts.
23-11-2007, 18:44 h
© Sebastian Kügler