KDEmod is a modular and tweaked version of the K Desktop Environment that has been optimised for ArchLinux. A few thoughts from Jan Mette and Johannes Schriewer.

KDEmod - Do You Have Your Own Modded KDE?

06 Nov 2007

20-10-2020 // This interview appeared in October 2007 on Gazeta Linux (an
electronic romanian Linux magazine) website.
30-05-2010 // Sadly, Jan Mette has passed away. RIP Jan.

Q1. Please, in a few lines, introduce yourself.

funkyou  Hi, my name is Jan Mette, i am 29 years old and living in the heart of Berlin. I am an IT-specialist and currently working as a freelancer in various projects on an on/off basis, which gives me some time for free software, although i am currently looking for a real/decent/creative(tm) job. I have way too much interests to tell here about, but one of it is obviously KDE and free software ;-)

dunkelstern  Well my real name is Johannes Schriewer I am 23 years old and I am working for a software company in Germany. I got to the KDEmod project by submitting the execute feedback patch to funkyou and then chatting about his future plans for KDEmod. He mentioned the wish of some users to have a 64Bit version and because I had intentions to switch for 64Bit. I offered help.

2. Describe your Linux history. Why so addicted to Linus own kernel and oss software?

dunkelstern  I started using Linux about the release of RedHat 5.2 albeit not exclusively. That time I had many problems to even get it running :) I got really addicted to it at the release of RedHat 7.3 Valhalla, used it about 3 months and switched to Slackware after this. After using it some time I began distro hopping like everyone does in her/his linux career. I got stuck at Ubuntu or better Kubuntu for some time and finally found Arch<.

Well why I am so addicted to Linux… That’s a good question, but I think it is because the amount of control you have in using Open Source Software. I was fascinated of programming computers since I was 10 or something like that, I programmed DOS (primarily GFA-Basic), Windows 3.11 (VB3) and then around the release of Windows 98 I switched to Visual Basic 6. I finally outgrew it using hacks to get Assembler code into the fully managed VB6 code, so I searched for new challenges and tried Linux. Here you have the opportunity to not only learn from textbooks but from real code that managed to get into production. So I think I use Linux because I am curious how computers work.

funkyou  I am using computers since i was 10 or so. When i got my first PC after many years of Commodore old school goodness, it came with Windows preinstalled (obviously!), but it got a friendly neighbour after some time - RedHat 4.x if i remember correctly. This was the start of my Linux activities, and some years later (where i had to learn a lot) i replaced Windows completely. I used Debian for many years, then Kubuntu for about one year and then i discovered Arch and felt immediately in love with it :-)

I think FOSS is very important nowadays, as we are living in times where we are almost controlled by big corporations and their interests - at least they try it everywhere, just look around. FOSS gives people the ability to learn and do things theirselves and in a community for free, and that is something i can only support as it gives the power back to the people. Knowledge is the key, and you can get it all for free, which is especially important for emerging countries where people cannot afford expensive technologies… Why Linux? Well, it gives me the most features and flexibility out of all *nixes with the slight cost that it is always in-flux, so things can change very fast, but once you get used to it its no problem.

3. Why Archlinux ? There are plenty of linux distros around.


  1. It’s fast, simple, bleeding-edge and has a rolling release system, so no reinstall or upgrade- mess anymore when there is a new version…
  2. It doesn’t get in my way. There are nearly no limits, as you can customise your system as you like and not the other way around as i experienced with many other distros - they all have their limits.
  3. Ever created a valid package in under 2 minutes? Well, no problem with Arch.
  4. The community is the best i have discovered so far, always friendly and very skilful - you all just rock :-)
  5. Speaking of community: No other distro allows you to contribute so easily as Arch. You can install the system and start contributing immediately, and this is the biggest point in FOSS for me, the community and people working together on a shared thing…

dunkelstern  Archlinux because of pacman, dead simple.

I tried all major and some minor distributions and all have their problems with package management. Slackware has no real package management as I used it (the community had no place to collect custom built packages at that time), so you were stuck with compiling things on your own. RPM is slow and bulky and some people tried to unify Suse and Redhat and as an end effect packages ran neither on the one nor on the other distribution. Debian’s packages are beautiful with their apt-get manager, but they are a pain to build (I owned a repository for Amarok SVN versions for Ubuntu once, so I know of deb package building well). I wanted to try Arklinux when I downloaded Arch by mistake, and now I am stuck, never tried Ark though.

4. Why KDE? Because KDE is developed mostly by german hackers? SuSE strong support and development history?

dunkelstern  I am using KDE since it’s beginning because it was the first real desktop environment, Gnome followed later. I tried Gnome and other environments numerous times (Once I was stuck with E16 because of the eye candy. It could do semi transparent windows on my 166 MHz Intel system, I was fascinated). Later on, when I was using Slackware, I tried to upgrade the desktop environments to the latest release as the development cycle of the distributions wasn’t so tight like it is today.

Firstly I got KDE running and tried to do Gnome from this base. I became desperate trying to solve the dependency hell that Gnome is. And it got no better until this day. Gnome people implement things and then try to push it into the base system (like dbus, hal, etc.) totally ignoring that it isn’t possibly the best solution what they have coded. Furthermore Gnome is more split up and they try to use libs that are in place somewhere. That’s not a bad thing but it results in bad behaviour (‘Not invented here’ syndrome) and multiple layers of wrappers around these libs. In the end no one wants to maintain the mess. Furthermore they simply ignore things (sometimes it’s even the user) like network transparency. Just try to run a Cairo based application via X11 forwarding on a remote display. Have fun staring at the screen while it is drawn.

Long story short: KDE has the more modern base technology and hears its users. Look at the Gentoo forums and wiki to see the real dynamic of KDE. By the way, I did never like Suse, they hide too much from the user (in former times yast was no open source software, just to mention it), so their support of KDE was no starter for my love to it.

funkyou  KDE gives me the features and flexibility that i need while it is also performant and pretty. I am so used to it after the years that i simply cant stand other environments anymore ;-) Its just a matter of what technology fits best to me and my needs, and KDE never really disappointed me so far…

5. Why KDEmod ? Arch’s vanilla KDE package lack of split was the only reason ?

funkyou  You are right, the initial reason was the lack of split packages and also the numerous patches floating around for KDE, and i wanted to try them all and see which of them i can use. Yes, its that easy :-)

dunkelstern  Well partly. At first I wanted to see if KDEmod is better than vanilla KDE and I saw my opportunity to get my patches to a broader audience than just posting them to or the mailing lists. Tested patches are more likely to be accepted into mainstream KDE, so I thought of it as a testbed and playground. I didn’t realise how big the audience really is in the first place. I never tried to get the patches into vanilla KDE because the maintainers of KDE have other packages to maintain too on Archlinux because there are not many devs around there.

Secondly the Arch way is KISS and so they patch only what is necessary so my chances would be small.

6. Tell us about KDEmod history, start-up and initial purposes. A few words about KDEmod website community, please.

dunkelstern  History telling is more a thing of funkyou, he started the thing. I already talked about my initial purposes in question number 5, so I’ll talk about the community only now. First to say: The community is great, without the active people in the bbs we wouldn’t be here anymore. And as I see it the community around KDEmod gets its own dynamic currently, we don’t have to answer all questions anymore because people are so kind to jump in when they know something. Of course we would like to see some people supporting us with additional patches or special GUI tools development. I finally got some Jabber chat buddies from it and it’s really fun hanging around with them online.

I enjoy the community, why not be part of it?

funkyou  When i started KDEmod, i only added some patches and built packages for myself. Basically it was all about learning how to get the most out of KDE on Arch and how to create packages, edit patches and such stuff…

Then i tried to split the packages, got it working and eventually posted the PKGBUILD’s and a repo with a small amount of split packages at the Arch Forums, because there were always people asking for a split KDE. This grew with the time, as people asked for more split packages and also did some packages for theirselves and sent them to me for inclusion into the repo<.

After some time, i created a small website with installation instructions, then Dunkelstern came along and offered his help and efforts. At this point there was no Forum or something you can call community, just one thread at the Arch Forums… This thread got bigger and bigger and also more confusing, so we created a forum + bug tracker + websvn and this was a success, as it all got much more organised. As you can see, the whole thing grew and grows with the people around it…

Well, Dunkelstern has already written about the community, and i cannot add much to it beside that its really fun to interact with the users and get stuff working better.

7. Your KDEmod daily routine. Maintenance, coding, packaging, testing, releasing.

funkyou  I check the forums and the bug tracker at least one or two times per day. When there is a new post at the forums or a new entry in the bug tracker, i get a mail, so i answer/react mostly quick… Once in a while i check for new patches and stuff and discuss about them with Dunkelstern. If they are ok and working stable for us, we include them. If there are bigger changes, we put the packages into our testing repo at first to let the users check out and report back, it is almost the same as in Arch. This has not always been successful as we broke KDE more than once very badly, but we do our best to provide quality packages because we are also using it daily. When there is a new KDE release available, things get sometimes a little bit hectic and we have to adapt and edit some patches, but so far it always worked out good…

By now, there is even an Arch + KDEmod installation working hard at the deepest Jungle of Sulawesi/ Indonesia, as my girlfriend it currently there doing her doctoral thesis, and it is doing a good and reliable job :-) Then there is of course the maintenance. I try to reserve one, sometimes two days per week for packaging and bug fixing purposes and also deliver new packages instantly if there is any big breakage. This is a hobby /enthusiast project, so please forgive us if there are any breakages from time to time - but we are also learning fast how to deal with it and it gets easier from release to release.

dunkelstern  My daily routine, let me think of it. Well I am very interested in KDE development, so it is not some kind of routine. If I find a patch on the net I look at it and if it is not too hack’ish, I just integrate it for testing on my machine. Makepkg does the rest :)

I work with KDEmod everyday, so testing is a matter of just using it. If there are known problems of earlier releases I test the thing just for regression. The bug tracker of KDEmod is a good help here. Releasing is a bit of stress currently because my internet connection is not as stable as I would want to have it. So uploads to the KDEmod server disconnect often and I have to restart them.

Shortly before a release there is much chatting with funkyou about errors we found or porting fixes. It is an ongoing development and it gets easier every release. Sometimes some errors slip through, so I have to apologise for it. But hey people you can talk to us, I get an email for every bug tracker entry that is made and I check back into the forums at least once a day. But please be patient, we get no money for this and it is a free time activity maintaining KDEmod, so bug fixes may last some days.

8. You have accomplished a lot of things with continuous development of KDEmod. Different hacks integrations, a release plan strong related to official KDE releases, powerful x86_64 branch, community growth, more and more split packages with each release. Think you could do it more? Or better ? Please, describe most important results of your project during time.

dunkelstern  At first, thanks for the compliments. Secondly, we are somehow at a limit of maintaining packages, therefore there is no current kofficemod however it was requested by users.

How to do it better? Well that would only work if we could put more time into it I think. Important results… how do you mean that? I try to tell you some of the benefits I can see. At first it is really fun to support such a big community and it is fun to read about what people do with KDEmod all the day. Well it could even be a good reference if I wanted to change jobs. It is not some kind of “project” you do and then ignore, it’s ongoing development and it is really interesting to get to know the inner workings of KDE. I learned a lot while hacking on the source. Even how not to do things (see hwinfo and kickoff code for that).

funkyou  I also have to say thanks at first. Dunkelstern already answered a lot here, but i will add some more of my sights about this. At first, we both are KDE enthusiasts. Dunkelstern more from the point of a developer and myself more from the point of a user / packager / tweaker, so if there is a new release of KDE available, we want it, as fast as possible :-)

How we could do more? Well, i think we already have reached a healthy point in terms of splitting packages. Splitting it up more would only cause more work and probably confusion from the users. In my opinion, nobody needs a split kdebase for example… In terms of patches, i think we have included nearly every usable patch that is available for KDE / Qt, so there is no more material to work with so far…

How we could do it better? Well, thats a tough question… At first, time is the key factor here, as Dunkelstern already pointed out. I think, with more time we could do it a lot better. As said before, we have broken KDE very badly for one or two times, and this is something i always want to avoid at all cost, which is sometimes a little bit difficult with all those patches and extra stuff that is included. For example, we recently noticed a breakage within our theme stuff and i guess with more time and more testing this wouldn’t have happened. There is definitely room for improvements, and we are already trying to do the whole thing in a more systematic and clean way. Also, we need to think more about our moves and the general direction. We included a new default theme with KDE 3.5.7 for example, but nowadays i would rather provide the theme as an optional thing than making it the default, because its still K.I.S.S. you know, and maybe this will happen with a later update…

Its all about how we can accomplish the best solution, and i think we will get better in this term from release to release, as time has already proved… About the important results: For me it was definitely the point where we had almost all KDE packages split, because that was also the point where more and more users started using it. Then there was also Dunkelstern’s makepkg modification that decreased our workload from over 200 PKGBUILD’s down to about 40 by adding the possibility to pacman to create more than one package out of one PKGBUILD…

The next big thing will probably be KDE 4, and from what i have seen in my weekly SVN builds, it will just rock :-) and also enable us to do much better and cleaner packages and splitting…

9. I think that last release of vanilla kde (3.5.8) will be the last for 3.x branch. Your KDEmod future plans regarding 3.5.x series? Just maintenance and bug correction releases and small patches for the future ? What you gonna do when KDE 4.0 is officially adopted in ‘testing’ and then ‘extra’ repos of Archlinux? It is possible to maintain an Archlinux new distro fork, containing KDE 3.5.x (precisely KDEmod) - let say a live CD ?

funkyou  At first, there is every indication that there will be a KDE 3.5.9 release and possibly even more, just take a look into the 3.5 branch in KDE’s SVN. Our plan so far is to provide the stable and proven KDE 3 in our current repo until KDE 4 is mature enough to replace it without regressions. Then we will see if there are people who want to maintain KDE 3 further (if they need it) and give the stick over to them… As always, we will provide a split KDE that conflicts with the one available in the official Arch repos, and we will provide a split KDE 4 in an extra repo once its out, but i will not be our main “distribution” at this point. However, time will show how this all will progress… When there is time and the possibility, we maybe will do a liveCD with KDEmod in the future, also maybe with some tools for administration/packaging, but is is something far in the future, as we are only 2 Devs so far…

Forking Arch is and has never been a real option, because i think its already perfect like it is, and i also think that providing extra tools to make maintenance easy and trivial is not a good step because it hinders people from learning how to really use and control their systems, and it also creates its own breed of problems to deal with…

dunkelstern  I have to say I read thoughts of supporting KDE 3 after the release of KDE 4 , so a 3.5.9 is very likely. Future plans are making it faster and even better. If we can’t find patches on the net anymore we will start to roll our own as it has been the case for many things new in the 3.5.8 release. When KDE 4 stabilises we will try to support some kind of KDE testing repository, but no promises when that happens and what gets into it.

We never planned to fork Archlinux, we would need about two dozen developers for that and we don’t have them :) A live CD is in the heads of funkyou and myself some time. If we get all bugs out of KDEmod we will definitely try to do that.

10. KDEmod project is getting bigger. You need any specific help ?

dunkelstern  Help would be appreciated. Specially for GUI configuration tools that Arch currently misses. We don’t want to make Arch more newbie friendly but we want to have some tools to ease work with it by people who know what they are doing. I mean I am not afraid of text configuration, but clicking on a button is simpler than reading the 100 pages manpage.

Core releases in 32 and 64 Bit work good as they are currently. If you would throw more devs on that site you only get more problems as communication will be more complex and enhancements would need more time as it has to be more coordinated then. But we would like to have some people that are not afraid of breaking their system by testing new releases of KDEmod before they get to current. I mean for sure we won’t release stuff that doesn’t work, because the first station in testing is our own machines. Furthermore it is possible to port KDEmod to other distros, if someone wants to do this send mails :)

funkyou  Of course we could need some help, as Dunkelstern already pointed out. Packaging is not a big problem and we are already very experienced with it and have a good workflow, but we would like to see some additional graphical tools for administration and package management, as there are none available yet for Arch / KDE … Also, i would like to see more feedback from the users if there are problems, especially more detailed reports. This is something i always stress about :-)

To make it short, you can help us by looking at our Forums and see if you can help with developing graphical tools, it all just started… Also, new patches are always welcome. If you have written a patch for KDE, just contact us and we will test your patch and if it runs ok, then it will be included into KDEmod. So far we were always fast in terms of including unofficial patches, and i don’t think that this will change at any point.

And at last, one thing. If you have a Web server and want to help us out, please consider to make a mirror of our repository. It is currently located on one lonely server in Germany, and although its a beast of a machine with a fast network connection, it sometimes gets a little bit over-crowded, especially when there is a new release or fat update.

To speak in numbers: Our repository currently generates more than 500GB traffic per month, and this increases steadily. Also, not everyone can access the repo as fast a we can, like people from overseas… If you would like to create a mirror of KDEmod, just contact one of us and we will provide you the details how to accomplish that. Thanks in advance.

11. A few words for Romanian KDEmod users, Archlinux users and why not, Linux and OSS software users.

dunkelstern  I’ll take the OSS route then and my words for the OSS users are: Get into the community, give something back, even if you only support other people or tell developers of spelling errors in their interfaces, all help is appreciated! Don’t be afraid to aid a developer finding a bug, even if it means installing a debug version and hacking some weird commands into a black window with white text, not really knowing what you do. Developers love it if you are so kind to help them and they will be open for your suggestions. Most people even tell you why and if they will accept your proposal or why they don’t think it’s a good idea, so both sides can learn from each other. We don’t have to convert all users to developers to be happy :) But teamwork is the key.

funkyou  This is, again, a point where i notice how small the world is. I never really thought about how many people are using our stuff or where they are located, but i come more and more to the conclusion that this is of course an important thing (note to self: phew, glad to see all KDE translations in our repo, finally), so i send my best greetings to our Romanian users, and of course all Romanian (Arch)Linux and FOSS users. My very best friend comes from Romania and if you all are like him, then you are definitely on my list of the very nice people :-)

Well, not much more to say. Using FOSS and contributing or being active in this community is definitely the best thing you can do in todays software-world, as all this is a totally unique and very worthwhile thing for us all. I am not just talking about ‘computer stuff’, its not just some sort of ‘licensing stuff’ or ‘gratis software’ - its a way, maybe even a lifestyle of thinking and interacting with other people and getting things done with fun, regardless of money or status or what not, and this is something worth to be protected, if not one of most worthy things we have nowadays…

Take care of yourselves and all the best!

Well, thank you for this opportunity.

Tags: linuxeng