However, so I thought, that either there just haven’t been many themes released since I’ve last checked (afair over a year ago) or there just haven’t been any good themes that were submitted and accepted at AGO. To verify that, I’ve taken a look at the gnome-look.org themes-section and proved my assumption true: I browsed through the first few pages of the GTK 2.x section and my eyes began to hurt. Then, I sorted the section to start with the highest-rated themes and my eyes hurt even more. One theme was – in the matter of quality and usability – worse than the other. Everywhere you looked only rough-cut pixmaps thrown together, added some really-not-looking-good background images to the menus and the window elements themselves and finished it all up with a foreground-color that either provided an exaggerated or an awfully low value of contrast. Meh.

The bummer is, that the majority of all themes look like that and only a few ones, mostly created by known artists like roberTO, Jakub and others really look tasteful and qualitatively good. In my opinion, this was and still is a major problem of the whole OpenSource community. OpenSource gives you the power to choose, modify and re-distribute, but I guess that exactly this power is being used in a wrong way – not only in the matter of control themes!

In general, especially within the Linux area, there are nearly no standards. Spoken from the designers’ view, there are not enough definitions like the GNOME UI-Design Guideline or the Tango Project, which try to convince and help the developers and/or designers to draw qualitatively better themes while still keeping up the freedom to choose and create. Of course, this won’t stop misbehaving designers to submit themes to un- or sloppy-moderated sites like gnome-look.org, but still it would provide the GNOME folks a “pressurizing medium” to say “You make it the good way, your theme might make it into our official project releases or at least on the cover of the official sites!“.

Though, for a real GNU/FSF-guy this way might sound like to much of “controlling” and “regulating” and by that lead to an operating system like for example Mac OS X (no, not Windows, there you have an even bigger problem regarding applications that look totally different than others) is. Still, most of these guys forget, that without at least a little bit of guide-lining, regulating and separating the wheat from the chaff especially the Linux Desktop will never make it into a higher market-share. There definitely is a reason, why companies like Novell and Red Hat keep up the hard and cost-intense work on their own UI-designs and improvements. If you want the users to be convinced about using a clean and stable operating system, you cannot simply stick with a UI on which the users’ thoughts are “Uh.” from the first click they’re doing. And of course, tastes are different and each user has a different one, but in one point all users will share the same opinion: An UI needs to be tidy and neat. No pixels. No exaggerated anti-aliasing (which should be better called “blur” in 90% of the existing GTK themes). Just a sleek and intuitive interface with clean structures and without distracting or even deranging elements (… like black backgrounds, white foregrounds and pixmaps that remind you of some white-noise-graph).

On software techniques the GNOME community seems to finally has understood what KDE is doing for years now already. There has to be a clean infrastructure (or backbone or whatever you’d like to call it) for solving problems and providing features. GNOME has started the move to GStreamer some years ago and now finally also moved to a backend (PulseAudio) which provides such an infrastructure. Also, introducing D-BUS and the HAL was a big step for the whole Desktop-Project, and the Gnome VFS seems to be trying to really compete with KDE’s now. So, as it seems, developers have finally recognized, that (especially in enterprise use) a desktop with no integration and where each application works different and uses a different infrastructure for providing audio, video or whatever else will never succeed against “the big ones”.

Unfortunately, in the matter of UI design, it still seems to take a while until contributors understand that it’s worthless drawing themes that look like Vista’s interface printed on a dot-matrix printer. By that, users of other desktop systems will always keep looking and thinking of Linux to be an unstable and totally not-integrative desktop-system, hacked together by some crazy, long-bearded freaks. Because for low-brown users, the UI is an essential element that helps them deciding whether a system looks usable to them or not. With an interface where each application looks the same, acts the same and allows the user to get this work done in an undisturbing way – and maybe adds a little bit of pleasure with smooth and clean looking effects (and by that I don’t think of wobbly windows!) – even someone that’s new to the matter will be able to get in touch with it quite fast.

I’m still waiting for the day on which especially GNOME’s interface-nazis finally make the move and decline all applications that do not strictly follow clearly defined designing guidelines for a clean and usable UI – even if it would throw out half of the applications shipped with a regular GNOME desktop (like Pidgin, OpenOffice.org, and so on…). Until then, I guess that Linux itself can be as solid as a rock, as fast as a lightening and free as free beer – it won’t be able to increase its popularity and climb the higher market-shares. Just because of the “look and feel”, which sometimes is just more important then pure functionality. Else, we would still be working on the CLI, wouldn’t we?

// btw: This is my 500th post I’ve been writing within over four years now already. Heh.

UPDATE: I just talked to gicmo, since he has Parallels installed on his Mac, and asked him to give me the ouput of ps aux | grep -i parall, so that I can compare it to my ps aux | grep -i vm (yepp, I’ve just installed the Fusion Demo) output… and well.. I guess I can call it equal. Each of these two products has some daemons running in background, even if the application itself isn’t or even hasn’t been running since the Mac started up. For me, both of them solve the actual problem I have: Virtualizing other operating systems (Linux, Win, UNIX, etc.) on my Mac. Now I need to figure out, what the details are, that make one product look better than the other. I guess I’ll need to search for some comparison of “Parallels vs. VMware Fusion“, using the recent versions of both. Maybe I’ll find something.

The actual reason, why I upgraded was KVM. Lenny seems to be getting more and more stable and by that usable as server system. Of course, it’s not etch yet, but since I’m trying to migrate from VMware (-Server) to KVM I’m actually forced to use Lenny. There are no KVM packages available for Etch, besides the ones from Backports.org. Unfortunatelly, even those are only available at version 28-4, what’s not that usable on a productive system.

However, so like I said, today I s/etch/lenny/g on my APT sources.list and ran an aptitude dist-upgrade. I was pretty surprised how fluent the upgrade worked out, the only thing it broke was the Roundcubemail I have installed on the machine for checking the Maildir via HTTP(S). Apache still works fine, Courier also runs, Fetchmail, and all the other stuff I had configured upgraded with nearly no problems.

Unfortunately, the upgrade still caused problems regarding my everyday-work. The most annoying thing is the fact, that I’m forced to use Icedove as mail client from now on, at least until some strange bug regarding the SSL authentication has been fixed in Evolution. The curious thing is, that with the previous version of Courier, Evolution worked just fine. I don’t know, what exactly the Courier developers changed, but it caused evolution to not be able anymore to authenticate against it. Some might say, it sounds like an Courier bug, but to be honest, I know how things are being implemented into Evolution and I don’t believe it’s Courier’s fault.

It’s pretty much a bummer, because I actually liked working with Evolution. Not because it’s fast – it’s not. And not because it’s light – it’s neither. And not even because it works perfectly in the infrastructure I’m sitting in – it definitely did not. But Evolution unfortunatelly is the only mail reader for the GNOME desktop environment, which really integrates into the desktop. I could always take a quick look at the meetings/appointments I have for a specific test, by just clicking the the clock on my menubar. Also, it was integrated into GNOME’s keyring, so that my mail-account password was just getting unlocked after the login.

However, this tiny, integrative features enhanced the usability of Evolution a lot, even if everything else might just suck. Now I’m trying to find out which bugreport documented the problem I’m experiencing, to put myself on CC. I’m really wondering when this will be fixed. If it takes as long as fixing problems regarding shared folders or caldav, then I guess I’ll just try to say Hello! to my future e-mail client Icedove.

Meh.

Altogether the whole system runs pretty smooth, unfortunatelly it doesn’t seem to be designed for mobile use, since it doesn’t just not implement “Locations” – it actually breaks the whole concept and does the exact opposite. It’s a bummer, because the first impression when seing openSuSE makes you think of it as a pretty cool enterprise system with stable software that just works. I don’t know yet whether I will keep openSuSE or move along to something else… or maybe I will just recover the DD-image of my broken Debian SID and hope for some upgrades that might fix the problems the last ones caused. And if that works I would probably never ever upgrade again, unless I’d have a whole copy of the running system which I could just play back if the upgrade procedure fails again.

By the way, at work I had dialogue with a colleague which blamed SID for the failure on upgrading. But like I wrote yesterday I also told him, that on an unstable distribution with latest software I do not expect the software to work 100% – my Epiphany kept constantly crashing since one of the last upgrades on SID, but it didn’t bother me too much, because I accept the fact that latest software also brings problems with it. My system got screwed because of the upgrade-procedure itself, and this is what I just can’t stand. I’m expecting even the unstablest system to provide an upgrade-procedure which does not break a running system by configuring the packages wrong and cause wrong permissions on files like /dev/null. Not even on Windows I ever experienced that the upgrade-procedure broke the system – it was always the things included in the service packs, heh.

However, to cut a long story short, I just informed myself about the pricings on the latest Apple hardware. Besides I’m really planning to get an iPhone + T-Mobile contract for around fifty bucks per month. It’s pretty expensive, but I think it’s worth. At least, it seems to work. Unlike OpenMoko, heh.

However, I’ve also been thinking of some new stuff I could implement into REX and had a nice idea regarding the output-volume. A microphone could be connected to the mainboard and placed somehwere in the center. This microphone records the sound and compares the signal-line with the one of the audio that is just being played. Depending on the grade of differences it could then automatically increase the output-volume to reach a acceptable signal-strength again. I got this idea when I became pretty annoyed of the fact that everytime when I open my window while driving I can’t hear the music anymore, especially when driving fast. By this feature REX would automatically adapt the sound to the actual soundscape. I don’t know how hard it is to implement this feature, since I’ve never done anything like sound-analytics in programming before, but I’ll find out soon.

Btw: And still I’m searching for some Map/Navigation software for Linux. It is pissing me off that nothing useful exists. I’ve found a project on BerliOS called Navit, but it doesn’t seem to be pretty far right now.

However, I’ve also been thinking of some new stuff I could implement into REX and had a nice idea regarding the output-volume. A microphone could be connected to the mainboard and placed somehwere in the center. This microphone records the sound and compares the signal-line with the one of the audio that is just being played. Depending on the grade of differences it could then automatically increase the output-volume to reach a acceptable signal-strength again. I got this idea when I became pretty annoyed of the fact that everytime when I open my window while driving I can’t hear the music anymore, especially when driving fast. By this feature REX would automatically adapt the sound to the actual soundscape. I don’t know how hard it is to implement this feature, since I’ve never done anything like sound-analytics in programming before, but I’ll find out soon.

Btw: And still I’m searching for some Map/Navigation software for Linux. It is pissing me off that nothing useful exists. I’ve found a project on BerliOS called Navit, but it doesn’t seem to be pretty far right now.

REX Car Entertainment System

To get the whole thing running you first of all need to get VirtualBox (not the OSE!) 1.5.6 and of course the REX.tar.gz. After creating a new Guest OS (Linux, Kernel 2.6) and adding the REX.vdi as Virtual Hard Drive you can already start it up. The first thing you’ll see will be the Splashy Bootsplash (which of course I’ve designed to fit into my car, heh). After some seconds you should see a regular Linux login-prompt. You can log into the system using root:install. After that you can start the REX user-interface by typing /etc/init.d/rex-ui start go. To make the interface start right after the boot just edit the init.d-script to remove the “go”-check. I didn’t make it start right after booting since I’ve experienced problems when doing that. Besides, the mouse does not work. Well, like I said, it’s a preview. Somewhy the Xserver doesn’t recognize the mouse so that you’ll need to navigate using the keyboard – but it shouldn’t be too hard.
However, altogether it’s like I said just a preview to make a first impression of how I’m imagining REX to be. I’ll package the software itself (REX User-Interface + Bootsplash Theme) as Debian Packages and provide them here for download so that everyone can build up a custom system.
Anyway, enjoy!

Download REX.tar.gz (Hendrik’s Mirror)

This means that I still won’t be able to provide a first preview of REX, what I’m really sorry for. Somehow it seems like everything just has to go wrong at the moment…

Btw: I’m still searching for a place where I could upload at least one GB. If anybody knows a good place, please feel free to mail me.

However, I have the system running virtualized on my 1.6 GHz Pentium(R) M processor and I noticed that the system-boot takes quite a lot – not because of the services, much more because of the whole hardware detection and module loading the kernel does. Though, I think that it would startup pretty fast when not running virtualized on such an CPU – I wouldn’t like to use some bloated Dual Core crap only to make the Enterteinment System boot faster, because for the actual work that REX needs to do a 1.6 GHz M should be more than enough.
I’m now trying to find some location on the web where I could upload the 1.7 GB VDI so that everyone can have a look at it – unfortunatelly the traffic it would produce when downloaded would be quite too much for my Webspace. If anybody of you does know a good place for uploading the package (besides SF.net or BerliOs): I would be thankful for some hints.