In the land of KDE…

Okay, I ended up with the cool titles for blog entries, I know. This entry is supposed to be a mix of stuff related to KDE, thus the name.

Let’s talk a bit about KDE, that is my desktop environment of choice (I’m one of the lovers of the “full-fledged environment”, so I don’t like minimal stuff like FluxBox and similar; what some people call blot, I call integration, and I like it), and that should have been clear for a while now. Together with Blind Guardian, KDE is probably one of my favourite German products 🙂

First of all, I want to express my congratulations to Sebastian Trüg, author of K3b, that’s doing a pretty neat job, creating one of the most complete burning interfaces I ever seen. His contribution to the Linux desktop is way more useful than the continue bitching of cdrtools author Shilling. If you can stand experimental software, I’d suggest you to try the 1.0_beta2 release currently in portage, it’s awesome.

Unfortunately, there’s a problem with this: this version, probably for solidity checks, disable the mediamanager service in KDE, and Amarok 1.4.2 and later uses the mediamanager for the “dynamic collection” setup; if you have Amarok open while burning, you end up with the collection disabled and a lot of features missing… don’t worry, just close and reopen it and it works. I’m sure the Amarok team is going to fix this problem before K3b 1.0 is final, or at least they’ll try… right Mark 🙂

While I’m really liking most of KDE development, and I really hope that the new #kde-ruby channel on freenode will bring more Ruby developers on writing KDE applications, as thanks to Richard Dale it’s possible to actually have perfectly integrated KDE applications written in Ruby rather than C++ (which allows a quite different approach to the whole development, gone the dead time needed to build, or rather to link, the code), there are a few things that I still don’t like, and I think would need a better development, or actually some kind of development. Unfortunately my time constrain don’t allow me to try to improve the situation, but I put my hopes in KDE 4 🙂

First of all, the KRDC/KRFB pair, client and server for desktop sharing via VNC protocol (and frontend for rdesktop, RDP client), seems to me like they are ignored a bit too much. The client does not work that well with xinerama, it loads full screen automatically, but appears only on one of the screen, sized like it was stretched in two, so you see only half o the window, you have to reduce it to window to use it on a xinerama-enabled setup. The server instead, using a very old version of libvncserver, and missing support for most of the recent interesting features of XOrg, it’s slow to the ridiculous when compared to something not that different like x11vnc. The only good thing the two of them have is the service discovery support, but they don’t seem to support zeroconf/mdns protocol, that seems to be what most of the software is standardising toward (thanks also to Lennart’s Avahi – that I criticised and still criticise from time to time but it’s proving the most stable between the choices we currently have, too bad DBus is still behaving strangely on FreeBSD without being patched and thus we can’t use it there yet :/); similarly, x11vnc does supports neither of them, which loses one feature, but not a core one (speed beats service discovery).

Then I still don’t feel that comfortable with Kopete; for some reason the latest update broke MSN filetransfer for me at least, and KMail doesn’t seem to improve on the memory usage side nor on the crashy side… Plus, I don’t really like their way to handle bugs, I submitted one once (stating that some e-mail addresses were incorrectly reported as wrong when they were correct, just strangely formatted), and providing a patch that rewrote the address parser to use regular expressions rather than manual parsing, as a try… when you get told to fix their code, and not rewrite it, if you want the bug fixed, you can be a bit pissed off, I think (especially since they didn’t fix the issue, it’s still there).

So my bottom line is, that I hope KDE4 will try to focus on fixing these problems, rather than going on with “usability audits” (that often result in stuff nobody but the usability people like) or with Windows portability (fine if someone wants to take care to port it, not fine when you hinder the development to re-focus it on Windows portability). I really look forward for KDE4, cmake or not, and I really hope these issues can be fixed.. or that with Korundum we can write a better mail client than KMail 😉 Maybe writing Tinymail Ruby bindings would help…