aKademy Interview: Developers From Kalyxo Project

At aKademy a talk about Kalyxo was given by Kévin Ottens and Peter Rockai. Curious what this effort is all about we
got in touch with the Kévin and Peter and they explained it all.

Please introduce yourself and what's your role in KDE project

Kévin Ottens: My name is Kévin Ottens (known as ervin on IRC).
I'm working mostly around the ioslaves in KDE (lately the new trash:/ and filesystems:/ ioslaves).
I also made some bugfixing and improvements to Konqueror and worked
on fuse-kio which is now hosted on

Peter Rockai: My name is Peter Rockai (known as mornfall on IRC). I work on Kapture which is an APT fronted for KDE.
I have lightly touched kwin in the past (and plan to revisit it soon).
Other KDE related activities are fixing bugs from time to time or implement small features
here and there as I see fit.

At aKademy you guys did a talk about a project called Kalyxo. What is this project all about?

Kévin Ottens: The primary focus is on creating and bringing innovative technology to the desktop user.
[To make the desktop experience smooth for the user, we are working on integration
projects as an important part of our mission.]
We focus on several areas, from desktop integration (among various toolkits and environments
like GTK and Qt) to system integration, where KDE integrates seamlessly with hardware and
other lower level features of the system.

How do you envision this system and desktop integration?

Peter Rockai: In the ideal case, the user would not need to worry about having GTK and Qt applications on his desktop.
The big part of this is the KIO-fuse bridge, which enables usage KDE network-transparency
in non-KDE applications. If implemented properly within the system (we are working on it), it should be
possible to edit images with Gimp over ftp without any hurdles (just like you can edit text with Kate
now). Same for and other 3rd party applications.

Kévin Ottens: We're working on the common look and feel too, and that's why we package the qt-gtk-engine
for the Debian users. But the real challenge is in the system integration, so that things
work seamlessly from the user point of view.

The future filesystems:/ ioslave is a part of this plan it will replace the old
devices:/ ioslave and go even further. We will introduce a way to do more guessing for
devices on supported platforms. This way KDE will still be portable which is a big plus of the project.
But we can provide extra functionality on specific platforms like Linux.

Since we will package some specific hotplug scripts for Kalyxo, we'll ensure that this
integration will work even better for Debian.

Peter Rockai: On another front of system integration, we are trying to close the gap between the underlying
system and its UNIX-inherited configuration system and the modern user interface KDE provides.
This part of effort is really specific to Debian based systems, as we need intimate knowledge of the
underlying system to be able to fullfill all the criteria users expect in such software.

So Kalyxo is sort of subproject and not another distribution?

Peter Rockai: Well, there are several aspects to Kalyxo. Currently, we are mostly focusing on the
augmentation of the existing system, yes. But for the future, we plan to spin off a new
platform project, which should be tightly based on current, preexisting Debian technology
and infrastructure.

Kévin Ottens: I would say it comes from two facts. First we're KDE developers so we contribute to it.
Second the structure of the Debian project itself. As soon as you work with Debian, you're
like a subpart of the project. But that does not mean we won't provides CD of course.

Is there any software or packages for people to try?

Peter Rockai: Yes, currently Kalyxo team is maintaining several KDE applications, some of them already in
Debian Unstable, rest in Kalyxo archive. You can find details on the archive location
on our homepage As for examples, there is
aKregator, the RSS
feed agregator, maintained by Pierre Habouzit in Debian Unstable.

Any Debian user can already start to use KDebconf (the KDE Debconf frontend)
which has been developed by Kalyxo and is now provided by official Debian packages.
Mario Bensi is currently working on improving its user interface and visual appeal
so we expect to have a new release of it ready for integrating into Debian soon.

Then, there is amaroK, the new media player
you may have heard about, which is currently in unstable branch of the Kalyxo archive.

Another technology you hear about much lately is NX and Kalyxo did not miss this either,
we now provide the FreeNX server in our repository.

Recently Pedro Jurado Maqueda joined us and maintaining Kiosktool and KGeography.
He is very interested by educational packages recently, which is a good thing for
sharing with a project like Skolelinux.
We're looking forward to cooperate with the
Skolelinux crew in the future.

Finally we also have some more experimental packages for inhouse projects like Kapture.

Are you paid to work on Kalyxo?

Kévin Ottens: As far as I know, nobody is currently paid to work on this project.

It is said that you guys are a reaction to UserLinux, the effort which
Bruce Perens started. Is this true and if so can you tell me why?

Kévin Ottens: Yes, UserLinux is part of our history. We're the children of the early
KDE-Debian effort (hence the name of our mailing list). With UserLinux
we expected to be able to work on a well integrated platform. Unfortunately
Bruce Perens chose to support only Gnome. Since then some people left this
effort very disappointed.

It was the dark age of the Kalyxo project and now we have this new cute name
and new plans!

Peter Rockai: I have to agree with Kevin on this. The exclusion of KDE from UserLinux was
pretty unfortunate event, at first. But then, we think it led to rise of Kalyxo,
which we deem a good end to the whole UserLinux and KDE story.

If the Kalyxo project tries to be Debian compatible then why not join the Debian

Kévin Ottens: In some way we're already joining the Debian project. As I said when you work
with the Debian project you're a subpart of it. But not fully joining it
allows us to be a bridge between the KDE and the Debian projects hence our
slogan "linking the worlds".

Peter Rockai: Another good point in refraining from being fully absorbed by Debian is the very
high entry barrier for new contributors. It often takes months before your first
contribution gets accepted into Debian and then usually even longer to become
an official Debian Developer.

We are trying to raise a more friendly and more open community in Kalyxo, to help
newbies contribute more efficiently and with more satisfactory results. By effectively
"holding their hands" while they create first packages, we let them learn by example,
instead of forcing them to read loads of technical documentation upfront. Our packaging
team will then check and upload the packages into Debian. We do not require that all
our contributors are Debian Developers, but we still encourage everyone to become one.

It is rumored there is a some kind of LiveCD. If so where is this available?

Peter Rockai: At one point, yes, there was a Kalyxo LiveCD, based on credativ's distribution, made
by Andreas Mueller. The iso images are currently not available though, because of
intermittent server problems and server migration. We will make them available
again at some point, but they are currently outdated and need more work.

What have you guys been working on at aKademy? Anything cooking?

Kévin Ottens: KDebconf got really more attention during aKademy thanks to Mario Bensi, he really
made a great job on it.

For me, aKademy has been a great opportunity to discuss the system integration
area with people here. Since some of them work on other distributions than Debian
it's interesting to improve the current situation and keeping the portability as
well. In most case, that mean that our design will be flexible enough to evolve
and support new technologies. Who knows which facilities the Linux kernel will
offer us in this area in a year?

Peter Rockai: Yes, we met lots of people from different projects, made new acquaintances and
of course had lots of fun hacking and talking with other developers. We also gave
a talk about Kalyxo at the user conference. We got some positive feedback on this
and a very interesting offer from a Spanish Debian-based distribution called Tuxum,
with which we hope to establish close cooperation soon. During the hacking week, we
spoke with few Skolelinux developers, and they support our goals and are open for
cooperation. We hope to utilize much of these new contacts to good of all the
involved projects.

What does the Kalyxo project need most now?

Kévin Ottens: Just like most free software projects - contributors! We want you! =)
Since we work on several areas we have three kind of tasks to offer :

  1. Packaging new KDE related software for Debian unstable, since we
    want that all the packages made by the Kalyxo project will enter Sid.
    It's a strong point of our policy, and we'll try to give advice and
    train you if necessary.
  2. Developing the missing KDE tools for Debian. We already mentioned
    KDebconf and Kapture, which would need more contributors of course.
    We're planning to work on system configuration tools too, we've some
    plans and early design discussed. But this is a task that will
    require more developers. System configuration means parsing, which
    means lot of work.
  3. And soon, we expect to work on our Kalyxo platform and CD. We'll
    snapshot a subset of Debian to provide it on a time based release
    schedule. But those snapshots will need to be maintained for bugfixes
    and security, we'll surely need help for this.


You two guys rock! (konqi also)
I wish you success! I also do hope SkoleLinux will be used in more and more schools, giving strong focus on the importance of Edutainment ( in KDE!

By a (maybe-not-so... at Mon, 2004/08/30 - 5:00am

Anybody who does work with Konq and ioslaves is okay in my book. Keep up the good work, kevin!

By Mikhail Capone at Mon, 2004/08/30 - 5:00am

This all sounds very nice, but why has this to be done in a distribution-specific way?
I mean, it's a good thing to improve integration in Debian.
For the moment.
But this creates other problems again.
Maybe there will come up some tool for system administration.
Nice thing.
But will I be able to use it on Gentoo?
On Mandrake?
SuSE? Red Hat? (you name it)
My opinion is the new kio-slaves and all this integration with other toolkits/applications is fantastic.
The integration in a specific distribution is not.
It's a work that need to be done x times (insert any number greater than 1000 for the x).
Maybe we should take a look at the Linux Registry project.
I think this is a better long-term solution.
If this is adopted widely we wouldn't need to care about distribution-specific stuff anymore.

If I got something utterly wrong and my posting is nonsense - please feel free to tell me so :-)

By SegFault II. at Mon, 2004/08/30 - 5:00am


Debian is the Universal OS, or at least is supposed to be. It
has no company behind it, but many companies base products on
Debian, and even more should.

The Linux Registry is not ready. Some people want to have what
works for Debian now. In the future, this may become portable,
but so far, it's only vaporware. I believe it can never really
become portable before the LSB doesn't cover it and lots of
code is changed.

Yours, Kay

By Debian User at Mon, 2004/08/30 - 5:00am

One small issue.

IIUC, the LSB specifies Sys V type startup scripts and a subset of RPM for binary packages.


By James Richard Tyrer at Mon, 2004/08/30 - 5:00am

I'd say integrate all the features other distributions have in debian. actually there isnt much debian has not... debian isnt that strong with source code, but thats it, imho.
some work on auto-package and zero-install to integrate it in debian whould be appreciated i think.

debian tries to be a 'meta' distribution - and it succeeds, look at mephis, lindows, xandros and others. rpm sucks quite a bit, while you almost never have dependency problems with .deb's... I think its really a pitty gentoo was started instead of improving debian, although I also think it was a good thing, to shake up things a bit.

but anyway, I agree with you on the linux-registry point.

By superstoned at Mon, 2004/08/30 - 5:00am

I think Gentoo is an even better "meta" distribution.
But thats my poor opinion, not based on any technical reasons.
Just how I feel - so please feel free to ignore it ;-)

You say deb hasn't problems with dependencies.
I don't think that has anything to do with the package-format itself.
Try apt4rpm or urpmi and you won't have dependency-problems with rpm neither.

Back to linux-registry:
Maybe I'll add an interface to this thingy in my apps.
It's a nice thing and it needs support to get widely spread.

By SegFault II. at Mon, 2004/08/30 - 5:00am

you will likely have more impact that userlinux which if it is not dea will likely die soon.

By AntiGuru at Mon, 2004/08/30 - 5:00am

What do you mean? Do you think Debian will die? Or the code they are writing will be erased? there is no way this project is a waste of efforts.

By Amadeu at Mon, 2004/08/30 - 5:00am

I _think_ AntiGuro ment that Userlinux will die soon, and that Kalyxo will have more impact than Userlinux had.

By Andre Somers at Mon, 2004/08/30 - 5:00am

Maybe someone of the project still reads this, as I'd like to say thanks for all your work and the things to come. That's one of the projects that could really bring Debian / KDE to a whole new level.
Now I only wish there was some more cooperation between all the KDE or Desktop oriented Debian projects, or those based upong... MEPIS Linux, MunjoyLinux and Debian Desktop Distribution come to mind.

By Christoph Wiesen at Wed, 2004/09/01 - 5:00am

There's some cooperation between Kalyxo and DDD regarding the use of The Swirl snapshots as a foundation for Kalyxo releases. And about MEPIS and Munjoy, AFAIK they haven't shown any interest at all in cooperating.

By Qerub at Tue, 2004/09/14 - 5:00am