KDE and Business: Intevation Interview

In the first in a series of articles, KDE Dot News will cover businesses with past and present involvement that are vital to KDE's ongoing success. Today, we interview Bernhard Reiter and Bernhard Herzog of Intevation, a company that has long been helpful with KDE application development and quality assurance.

Past

Please introduce yourselves

Bernhard Reiter (ber): I co-own and manage Intevation together with my friends Jan-Oliver Wagner and Frank Koormann. The company allows me to spend work-time on general Free Software topics while it also is completely focused on Free Software.

I have been Vice-President of the FFII in the past and currently I coordinate the activities of the FSFE in Germany.

Also I am a member of the German Grass User Association and active within the German Free Software Business Association. My education is in Applied Systems Science (German University Diplom) and in Geography (Master of Science from University Wisconsin-Milwaukee).

Bernhard Herzog (bh): I'm an employee at Intevation. My job title is basically "Lead Developer".

My educational background is physics with a "Diplom" from a German university. I have been programming for more than 20 years now, starting with Basic and assembler on 1980s home computers, later C++ and C and now mostly Python and C.

In the Free Software community I am probably best known for my Free Software vector drawing program Skencil.

What were the initial reasons to begin Intevation and what were the initial goals?

ber: Being system scientists, the three founders could easily see that computers were becoming increasingly important, both for society and business productivity. Having endured a lot of bad software without apparent technical reasons, we understood that Free Software has advantages for users and producers alike.

Jan, Frank and I wanted to do better software and founded Intevation in late 1999. We had polished the idea in the years before. The aim was: provide service and software to customers that we would want to buy and use ourselves. To reach this goal, we completely focused on Free Software, new development methods, strategic consulting and usability design.

Because we all had software project experience in this area, our first projects were in the area of GIS and scientific modelling. We also founded Free GIS, to show that it is possible to strategically progress one area, by providing and overview.

How did you join Intevation?

bh: I had exchanged a few emails with Bernhard R because of some bug in Skencil in 1999, if I recall correctly. I later met him at the LinuxTag 2000 in Stuttgart. As it turned out, Intevation was looking for a software developer, I happened to be looking for a job and a few months later I was hired to be the first employee of Intevation.

At what point did Intevation begin involvement with KDE and what was the reasoning?

ber: In 2001 we won a tender by the German Federal Agency for IT-Security (BSI) to create a Free Software email solution that could do signatures and encryption according to their standards (an S/MIME subset). We teamed up with g10code and Klarävdalens Datakonsult for the bid and implemented this within GnuPG, mutt and KMail as this was what the BSI liked. This is documented on the Äypten webpage.

What were early results of KDE interaction and why did you continue with KDE?

ber: The project was a success and we got a couple of other contracts with the BSI and other customers afterwards. There was a followup on the Äypten project and the contracts which led us to create Kolab, which is an email and groupware solution with a native, smart KDE client.

What was the inception and growth of the KDE PIM meetings in Osnabrck.

ber: During the first contract for Kolab, we invited KDE PIM developers to Osnabrck to improve communication and the technical integration of Äypten and Kolab.

I remember some KDE developers which were not in our team being sceptical. It was very good that the BSI allowed us to talk to the developers during the implementation of the technology.

So the meeting actually had a schedule and additionally I offered to help as a moderator to best use the time close together talking, not just quietly hacking.

Some developers found it strange to "close the laptops" for the meeting part :)

Obviously the group liked the meeting, thus we did it again. And the third and fourth time we got approached by the KDE PIM team to provide space and bandwidth for a meeting, and we were glad doing this. While the first two meetings were to some extent paid by our contracts, the developers themself carried most of the costs for the last two summits.

And they find it very productive to talk together now.

Present

What's your current relationship and focus with the KDE community?

ber: We continue to maintain the Kolab and Äypten solutions for our customers, so whenever we need to fix a bug or add a feature in the KDE part, we are making sure that it gets back to the mainstream development. This means testing Kontact and its components a lot and verifying bug fixes.

bh: We also build binary packages of KDEPIM and related non-KDE components such as the GnuPG-based crypto backend for our customers and the GNU/Linux distributions they use. While the current KDE releases come with most of that out of the box, sometimes the customers need features that are not enabled by default or the version that comes with the distribution is a bit too old. These packages also see some quality control, especially the parts that are important for Kolab and Äypten, which is something that the distributor will likely not test as thoroughly, if at all.

How do you feel that your business benefits KDE and how does KDE benefit your business?

bh: As Bernhard said, if we make changes to Free Software we use we try to give them back to the project in question so that it will be included in their official versions. This is not always easy, but in the case of KDE it works pretty well, because such changes are made by the well-known KDE hackers at Klarävdalens so that the changes are done in a way acceptable to KDE from the start.

Also, our customers pay us for those changes. That means they are features and bug fixes that are really needed by businesses and other organisations. They give us an insight into what they need and how they work with e.g. Kontact. This will make KDE a better system.

What were the topics at this year's KDEPIM meeting in Osnabrck [*] and what did you hope to accomplish?

ber: The meeting is completely there for the KDEPIM developers. We hope that they enjoyed their stay in Osnabrck and reached their goals. One goal was a vision for the PIM framework in KDE 4.

In the end of course, everybody will profit from the higher quality software; and everybody includes Intevation.

Intevation is known for KDE PIM and for your help with quality assurance and testing. Are there other elements of Intevation work in the KDE community that should have more visibility?

ber: During the time I coordinated the Äypten and Kolab contracts, I have noticed that this work did not get much recognition within KDE. These contracts made it possible to put excellent design in practise, e.g. the new groupware approach by Martin Konold from erfrakon and a modular security backend by Werner Koch from g10code. A lot code work for KDE was done, e.g. by Till Adam, David Faure and Bo Thorsen from Klarävdalens Datakonsult. As coordinator I influenced what was happening. And I kept the project going which meant more value for KDE.

When seen from a wider angle: Intevation doing this contract is like building a bridge between corporate users and the KDE projects.

KDE is part of the Free Software community and Intevation does a lot to promote this community in general. We support and participate in quite a number of organisations. Early on I mentioned some. For example within the LinuxVerband we, among others, made sure that KDE remains an option for the Free Software business desktop. Also, Intevation's yearly award for achievements of Free Software at the University of Osnabrck went to Carsten Niehaus for the KDE application Kalzium, a few weeks ago.

Any thoughts on recent KDEPIM events such as OpenSync?

bh: Collaboration with other projects on infrastructure that is not KDE specific is always a good idea. Having recently tested KPilot with regard to syncs between various Palms and Kontact -- functionality that several of our Kolab customers want -- I can see that the basic features are there but there is clearly room for improvements.

Shared infrastructure like OpenSync can help with this because it will increase the number of people working on it with various devices.

Are you satisfied with your involvement with KDE to date? Are there things that you've learned for working with the Community? Things that you hope KDE has learned from working with giving companies like Intevation?

ber: Naturally, we would be happy to have more customers with KDE related work. Also we consider ourselves part of the Community of KDE and Free Software. The KDE project is unique in that it is quite huge and managed to produce a lot of good applications with C++, which was surprising to me. But as I learnt from David Faure, Qt helps a lot to make C++ nicer, like being more dynamic at runtime.

I think many people in KDE meanwhile realise that an interest of a company need not clash with interests of the project. Everybody needs to be honest and keep the communication going. It is a special skill of Intevation to work with the existing communities and at the same time reach results for customers.

Future

Does Intevation plan on future KDE involvement?

ber: As a service oriented company, we depend on our customers. We will continue to support the KDE Kolab client with Äypten functionality for quite a while, I think, as customers want our support and are satisfied with it.

With Äypten and Kolab, Intevation has shown that we can create innovative solutions with KDE and integrate the right partners. This qualifies us for other IT challenges. Somebody just has to contract us.

One specific project we aim for is a KDE Kolab Client on Windows. (If you are an organisation willing to fund this, let us know!)

Does Intevation have any special plans for applications or testing efforts on anticipated KDE 4.0 release?

ber: Not yet, as customers do not think about deploying KDE 4 at this stage.

What can KDE do to improve working with Intevation or other businesses?

ber: It is important to keep in mind when communicating with us or other companies we usually see different problems affecting hundreds of users.

So our feedback brings in a different perspective.

As I am basically selling KDE to some organisations, I can see high level arguments which are unrelated to code, but still are important for KDE, if it wants keep the critical mass as a mainstream desktop environment. KDE has gotten behind in corporate mindshare because company people had no one to talk to. Communication needs to be more structured for companies to interface with KDE efficiently.

I think that the idea of the technical working group is a good start.

bh: From a more technical point of view, especially testing and packaging, I can see two things that would make our job easier. One is the packaging of the translations, the other is quality control.

KDE has one repository for all translations for all parts of KDE. This invariably leads to distributors having one package for each language that contains the translations for all of the KDE packages. When we package kdepim we usually have to provide at least German translations and that means we have to produce a kde-i18n-de package tailored to the customers Linux distribution because it also has to contain the translations for all the other parts of KDE.

As for quality control, I would like to see KDE adopt more automatic tests. KDEPIM for instance contains some tests, but they cover only a small part of the functionality. During the Äypten and Kolab projects it happened occasionally that bugs that were already fixed were introduced again later. At least some of these could have been prevented with automatic tests. KDE seems to be improving in this area, though. The move to Qt4 could help with this, too, because it emphasises the Model-View-Controller pattern which helps to separate the program logic from the user interface.