Transcript of talk: " has happened, is happening, will happen"

KDE Contributor and Developer Conference

Notes taken by: Jonathan Riddell (these are not official material)

More information on the talk can be found here (hosted on have released 6.7 and will soon release 6.8, then X11R7, incredible given performance in the past few years. Also host modular X and DRI/Mesa.

They host various other projects. Cairo is a 2D vector graphics framework with multiple backends. D-BUS is a system wide IPC. GStreamer in a multimedia framework, HAL is the hardware abstraction layer.

KDrive is very shiny, it's a good codebase for proof of concepts such as composite rendering and a new acceleration architecture. Modularisation is a win for everyone. They have been looking at alternatives to CVS and are considering TLA which does good things for merging. Next may have composite and may move to /usr and be modular.

A lot of people believe X sucks. X doesn't suck, XLibs does. It tries to work for both application authors and widget kits and fails at both. XCB is a replacement, very low level, the X protocol in C. XCB is not suitable for applications but allows toolkits to hand tune and optimise their codepaths. Supports most of the X extensions. Getting rid of XTrans (the most evil spaghetti code ever - X's transport layer, it deals with everything by #include'ing a bunch of .c files).

Cairo is a vector graphics library, backend independent, exports to X and other places and is supported by GTK and Qt4.

D-BUS is the universal remote procedure call system. Separates system and per-session busses. Allows deep integration with the underlying system e.g. HAL. One idea would be to use HAL for a single network configuration e.g. IP address or HTTP proxy can be known by Konqueror without Konqueror knowing anything about the underlying system.

UIM and SCIM, Universal Input Method and Shared Common Input Method. Allows far better input of non-Latin character sets, X and non-ISO8859-1 character sets have been broken for ages. platform, first release in 10 days! Common base of libraries, specs etc for desktop developers, distributions and software companies to build upon. No D-BUS, HAL or Cairo (or system trace apparently) yet because it has to be API stable, but will have e.g. menu standard.

Services provided by are ssh, CVS, web hosting, SVN (WedDAV and SSH) Arch (WebDAV).

Help wanted, always lots of projects needing help and new projects are welcome. If you have a cool library which would be useful take it to freedesktop.

There is not much of a decision making process, if there isn't concensus then something else is wrong. Daniel is the current release manager.

Comment that the decisions for release management looks like Red Hat decisions. If anyone wants to act as a better release manager they are welcome to. Another comment that the platform release doesn't matter ultimately, if you don't like it don't use it.

Leon from comments that he was strongly in favor of the new using free desktop. They need more feedback and involvement from KDE. needs help to make sure they don't make mistakes with the future of Free Software on the desktop. There has for example been a problem with xprint but there arn't enough people to solve it.

Matthies adds to what Leon is saying, Havoc is very open to participation, nothing that comes from freedesktop can be a standard unless KDE and Gnome support it. If KDE does not support it there is no point in having it in freedesktop. There is a strong need for more KDE participation. You (KDE Developers) can get involved, it's very cool and a lot happens without KDE which isn't a good thing.

Freedesktop isn't about standards in an ISO or w3 sense, it's not someone dictating what must be done, it's about KDE and Gnome talking to each other and coming up with the right answer.

Complaint about Linux Standards Base wanting to include GTK when Qt can not be a standard. Aaron says this comes down to involvement, everyone should know about the freedesktop projects.

Matthais says if we remember a few years ago with e.g. the clipboard we had to use the existing platform. Now we can change the platform to make Klipper not a kludge (it continuously polls the clipboard).

Daniel says freedesktop is the single coolest thing we have had in years because it allows us all to talk together. E.g. people have been saying "d-bus sucks because..." including one saying d-bus sucks because there's no Qt bindings. The reason there are no Qt bindings is because nobody was there to write them. Matthais adds the Harl has now written the bindings so we should check them out.