Fosdem Argon7
2005 Edition Free and Open Source Software Developer's European Meeting






Interviews

2005-02-09 - Matthias Ettrich

KDE

An interview conducted by FOSDEM & theLinuxFR readers
FOSDEM - Please present yourself.

Matthias Ettrich - I was born 1972 in southern Germany, and later studied computer science in Tuebingen. During my studies I discovered GNU software and Linux, and wrote the first version of LyX. This positive and successful experience of initiating a little self-sustaining free software community made me brave enough to start the KDE project later. After my degree I moved to Oslo, Norway, to work for Trolltech, the creators of Qt. Currently I'm holding the position of VP of Engineering for the development tools side of the business. Although VP stands for "very pointy", I'm still able to actively participate in development work myself.


FOSDEM - In your point of view, how do you see the relationship between the KDE and GNOME community, but also FreeDesktop.Org?

Matthias Ettrich - If you mean "community" as in "developer community", there is a professional level of cooperation through freedesktop.org. While many KDE developers had some initial skepticism towards freedesktop.org, it has now gained broad acceptance. This is good. When it comes to "community" as in "user community" things look very different still. A typical GNOME user seems to avoid KDE applications as the devil avoids holy water. Vice versa, a typical KDE user tends to avoid Gtk+ or Gnome-based applications. This creates unhealthy pressure to clone any good idea that shows up in one camp, which in turns creates lesser friendly feelings towards each other. Still cooperation is important, because it allows 3rd parties to integrate with either desktop, and increasingly so.


FOSDEM - LyX 1.3.0 is now 2 years old, and since then it seems to be almost in maintenance-only mode. So, what is going on? Are there any plans for a future evolution, or do you simply consider that it is now finished and good as it already is?

Matthias Ettrich - I'm afraid, I haven't been involved with LyX in a very long time, so I'm really the wrong person to comment on that. As a Qt developer I wish LyX would consolidate their graphical user interface, hopefully with Qt 4. While big projects and companies might be able to afford to maintain their own GUI abstraction layer with different frontends, it puts a significant burden on the developers, and it ultimately slows down innovation and progress.


FOSDEM - Is XML a future evolution for the basic lyx format?

Matthias Ettrich - Had LyX been started today, it would very likely use an XML-based format, but whether this is reason enough to change now, I don't know. Had LyX been started today it might not even use LaTeX, but maybe lout.


FOSDEM - Has KDE ever thought about using more high level languages like C# * or python? If so, what was the conclusion?

Matthias Ettrich - KDE is not a centralized body, but a vibrant community of hundreds of talented individuals. That means there's hardly anything that wasn't thought about at some point. We do have developers that feel strongly about python and especially ruby, and that push heavily in that direction. The conclusion? As of today, no bigger applications have yet been written in those languages for KDE; and those started in these alternative languages switched to C++ at some point. Personally I like having RAD languages like python and ruby available, but I wouldn't want my core desktop applications to use them. C# and Java are different as they are more general purpose system programming languages than RAD languages. Not only in that respect are they very close to C++. Had there been a fast JIT-compiling open source JVM, then the Qt/Java native API would probably have been a lot more popular, and we would see more Java in KDE. C# is definitely interesting in the longer term thanks to the Mono project.


FOSDEM - Are there any plans for KDE 4 to colaborate more closely with Gnome to establish a common framework for things like IO, database access, multimedia, configuration, usual desktop services like file open dialog, print dialog, systray icon, etc

Matthias Ettrich - Most of the things you mentioned are covered through continuous efforts done at freedesktop.org. In the area of multimedia there's even more cooperation, so I guess the answer is yes.


FOSDEM - The KDE project is now almost 9 years old. What arethings you want to see in the project before his tenth birthday?

Matthias Ettrich - During the past years KDE has grown from a relatively small developer community to a large number of people in special interest groups. KDE really is a community of communities, a meta community, which works very well. The biggest issue today I see is in the common core upon which all these sub communities base their work: kdelibs, kdebase and the user interface guidelines. At present, I fail to see how we should actively move those modules forward to accommodate the needs of the entire project, and who has the authority and the mandate to do it. Within KDE's legal entity - the KDE e.V. - we have started to use low-overhead voting procedures for decisions like new memberships or the location of the next conference. My hope is that in the future we will also be able to authorize members to be maintainers or lead architects in order to move the platform forward. With the size that KDE has today, an anarchical structure creates too much resistance to turn the boat around, unless you can elect a captain. The struggle of the usability groups for simplification is one visible example, but there are more. KDE has all the right ideas, but whether we are able to realize them against our own resistance is a different matter. Once we have authorized maintainers that can speak for the community rather than only and obviously for themselves, it will be significantly easier for KDE to be an active voice in communities like X.org or fd.o, and to cooperate beneficially with commercial entities. Today we don't really know how to approach those things.


FOSDEM - KDE is the Desktop Environment that enables the users of other Operating Systems to move easily to the Linux world. Did this happen indirectly or was it one of the purposes when working on the project?

Matthias Ettrich - Indirectly. KDE's main purpose has always been to provide a smooth and friendly end-user experience appropriate even for those people who wouldn't label themselves `geeks'. Since most people get exposed to other, particularly mainstream, operating systems and desktop environments before they have a chance to try KDE, it's natural that we see more migration to KDE than we see users for whom KDE is their first desktop environment. And since previous experience influences what we perceive as user friendly, it is also natural that KDE follows and benefits from some of the paradigms created in the MS-Windows and Macintosh world. We arrive there indirectly, by listening to user requests, not following the direct route of simply copying existing systems.


FOSDEM - A lot of users ask themselves why you are still focusing on the KOffice project when a solid OpenOffice?.org already exists. What are your thoughts about that?

Matthias Ettrich - A little variety and competition is a good thing for everybody. An Open Source project is not a company. If a company is doing A and B at the same time, one could argue that if they didn't do A, then they had more people to focus on B. So A slows down B, and killing A could accelerate B. With Open Source, this isn't the case. Someone who is interested in writing office-type KDE applications in their spare time, is not necessarily interested in for example writing free video codecs for a new multimedia architecture. If you are interested in office-type applications, and want to learn more about how to write them, or have innovate ideas how to improve them, KOffice is probably a much better suited base to start from. While it is behind in pure functionality of the word processor or the spread sheet, it has applications that OpenOffice doesn't have, it has the better internal architecture, starts up quicker, compiles faster, and is an invaluable test bed for Qt and KDE technology. It's overall impact on KDE thus has been very beneficial, and continues to be beneficial. But there's also an important global impact: what good was a standard office format such as OASIS if only one office suite implemented it? KOffice plays an important role in actually making the standard a real standard.


FOSDEM - What is coming in Qt4 ? What you think should be the focus for KDE 4 and your opinion on which multimedia framework to use ?

Matthias Ettrich - This is three questions in one, and we are running out of space! Let me start with the last one: the best multimedia framework for KDE is the one that our multimedia developer agree on using. They know more about these things than I do, and it's a difficult decision that requires many insights and indepth knowledge about the topic. When you pick a fundamental technology like this for a project of KDE's size, there are many things to consider. Technical quality and potential is only one of many aspects. Other aspects are proven sustainability of the project, stable maintainership, acceptance and support by other projects, and modularity. With respect to Qt 4: we hope to deliver the best GUI toolkit possible, based on years of experience with a broad, active, and vocal customer and user base. For us, it's a major step forward in terms of programmer productivity and graphics capabilities. I believe you'll see that clearly when the first Qt 4 based apps appear. And last but not least, my top 3 favorite focus areas for KDE version 4: usability, usability, and usability.


FOSDEM - What are you expecting from FOSDEM ?

Matthias Ettrich - Inspiring talks with interesting people, fancy Belgium beer, and tasty French fries.

 








© FOSDEM 2003-2005 - powered by Argon7