Looking Back, Looking Forward: GaŽl Duval on Mandrake

Timothy R. Butler
Open for Business

September 09, 2002

It is September, and that means brilliantly colored leaves, cooler weather, and a new Mandrake Linux release. As the big day for Mandrake Linux 9.0 approaches, Open for Business's Timothy R. Butler talked with Mandrake co-founder GaŽl Duval about the company's past, present, and future.

Open for Business: You started out with Mandrake as a project [ http://www.indidea.org/gael/mandrake5.1-original-announce.txt ]. When you started, did you envision it crossing over into a corporation, or did you expect it to stay a non-commercial project?

GaŽl Duval: I didn't expect it to stay a non-commercial project for several reasons. The first reason was that I didn't have much experience in non-commercial collaborative Free Software projects at the time. So I had more of a vision of how it might become a commercial offering than how it could be developed into a "Debian-like" product.

The second reason was that in the summer of 1998 - when I launched the first version - there was a very strong demand from Linux resellers and users for Mandrake CDs. Before MandrakeSoft was created, I had an arrangement with a small company to sell copies of Mandrake; they made some money with that.

As a result, I think the commercial dynamics around Mandrake Linux, and the creation of MandrakeSoft, have been key factors for its development and long-term success. But as you know, Mandrake is much like a Free Software project that is financed by a commercial company. This approach makes great difference when compared to other Linux distributions!

OfB: On the same note, exactly how did you get started with creating a distribution? Obviously this was something that took a lot of time and planning.

GD: I initially planned to release my own Linux distribution in 1997, when I had a lot of free time since I was performing a type of civil service for my country. The project was called "NetOS" (which has nothing in common with recent projects of the same name). The final product was planned to be offered as two different versions: NetOS Server and NetOS Home. The project was based on Slackware, with the Open Look Window Manager used as the default desktop.

When I began to work on that project I was given a retail pack of a famous Linux distribution as a gift for having translated several Linux HOWTOs. This distribution had two powerful features compared to Slackware: It had a great installation procedure (at the time!), and also had a great package management system. At that same time, I discovered the first versions of KDE. Then I decided to delay my "NetOS" project, started to hack and recompile the installer, packaged KDE and other "friendly" software, then released the whole thing 8 months later, just after KDE 1.0 was released.

Duval never expected Mandrake to stay non-commercial.

OfB: In recent times Mandrake has made a pretty big deal out of its support for Free Software/Open Source. For instance, in your communique explaining why Mandrake would not join UnitedLinux, one of the reasons was its "non-free" nature. Considering that one of Mandrake's original purposes was to make available a non-free (at the time) desktop, when would you say this policy change took place?

GD: I appreciate you noticed such a paradox! :-) The license of KDE has been a long story. The main issue was that Qt's license, in 1997, wasn't considered truly Free Software. This was because Troll Tech wanted to keep control of the Qt development tree, so they didn't allow any modification of the source code without their acknowledgement. This was, in my opinion, a small issue which was made into a big story since:

  1. The Qt sources were available to everyone
  2. Anyone could compile them, in any desired manner
  3. Everyone could redistribute the Qt sources and binaries for free

My understanding of the story - which is strictly personal and not related to MandrakeSoft official position - is as follows: Red Hat initially planned to include KDE but they were afraid of including Qt because they feared Troll Tech could then change the license and request royalties from Red Hat. So Red Hat refused to include KDE, and heavily pushed GNOME (then in its embryonic form) over KDE (this isn't a criticism, I think they were certainly right after all).

So I made a pragmatic decision. First, I explained to everyone that the Qt License wasn't so bad and didn't prevent them from using Qt as if it was real Free Software. Secondly, I sent several emails to Troll Tech guys to try to convince them to change the Qt License so that it would be considered Free Software.

I never thought that including KDE/Qt in Mandrake was much a threat to Free Software. And Red Hat seemed to realize the same thing because in early 1999, they started to ship KDE!

Anyway since 1998 I became more and more convinced of the need for true Free Software. It's the most powerful weapon against Microsoft.

As for the rest of the story: Later - in June 2000 - Troll Tech asked me (among other people from various companies) why I thought the Qt License should be changed to real Free Software. I wrote an essay for them about the benefits of a business model based on Free Software. In September 2000 they released Qt in GPL. So our efforts weren't in vain.

OfB: MandrakeSoft has received a good amount of criticism for its policy of asking for donations and club memberships to keep the company afloat over the last year. Do you think there is any merit to these complaints?

GD: I understand such complaints, but we knew that an honest appeal to the community was the best way to make Mandrake Club a success -- so we did, and it has become a success. Unfortunately several people were under the impression that we were "begging" for money. The truth is, MandrakeClub - which was created on user's demand months before our appeal - is a service that delivers real benefits to its members, it is not a charity system.

The reasoning behind the Club is as follows: the development of a Linux distribution is very costly, but the final product is available for free. The revenue from selling boxes is very small and doesn't cover development costs, so we encourage our users to join the Club and receive special privileges if they want to support Mandrake. Most people understand this approach and I think it's a valid business model, at least partly.

For the future, we are thinking about a "Mandrake Foundation" which would be a non-profit organization that focuses on developing the Mandrake Linux distribution exclusively. It would be financed partly by Club memberships and/or donations and/or by a "Street performer"-like system, and partly by companies that make money with Mandrake products, including MandrakeSoft. We think this approach would be much clearer for everyone to understand, and would also provide a more secure future for the Mandrake Linux distribution. It would also help MandrakeSoft become a more successful and profitable company by cutting most of its development costs.

OfB: In the now famous March 11 message talking about the future of MandrakeSoft, there is a reference to the "'sins' of the previous management." Would you mind elaborating a bit on the meaning of this?

GD: For one year, we had a so-called "World Class Management" team that left us in a very bad financial situation, and engaged the company in ventures (such as e-learning) that we should never have been involved with. But that's all part of our history now, so I'd prefer to not dwell too much on that.

OfB: Speaking of which, recent MandrakeSoft financial releases [ http://www.mandrakesoft.com/company/investors/newsletter/sn020722 ] look increasingly positive. Are you satisfied with the current progress Mandrake has made financially? How is the future for your company looking at the moment?

GD: I'm very satisfied with our latest results which show that our efforts are starting to really pay off. The trend is very good and I hope it will continue this way. But until MandrakeSoft earns more money than it spends - "break-even" is planned for the end of this year - we are still in a difficult cash situation. This is why we are currently conducting an Increase of Capital [ http://www.mandrakesoft.com/company/investors/bsa ].

My personal opinion is that Linux is getting bigger and bigger, and as we see more and more big names in the software industry coming to us, MandrakeSoft's future should be bright as soon as we become financially secure.

OfB: In a 1999 interview [ http://old.lwn.net/1999/features/GaelDuval/ ] with Linux Weekly News, you said that "Redhat is the IBM of the Linux market, we expect to be the new Compaq." Is that where you see MandrakeSoft as being today?

GD: That is quite an old interview :-) Maybe that's true, I don't know. Our chief goal is to provide an operating system that is so flexible and powerful that it can be used either for setting up a cluster for intensive calculations, or for playing games. Nowadays, Mandrake Linux is in the "top 5" of Linux distributions with partners such as HP, IBM, AMD, and other big players. Who would have bet on that three years ago?

OfB: What parts of MandrakeSoft are you involved in at the moment? Are you involved directly with the direction of Mandrake Linux?

GD: Yes, I'm involved with MandrakeSoft's direction, and also a member of MandrakeSoft's board. But I'm not a real "executive". One reason is because I've worked at home since the beginning of the company. This was my only condition when creating MandrakeSoft! On the executive side, I'm very involved in everything that is related to Web and Internet here, including communication activities and new projects (MandrakeStore, MandrakeClub, etc.). My biggest regret is that I had to stop developing when our development team became so large. It's not easy to manage a team when you aren't located at the same place as everyone else.

OfB: How has the initial reaction to Microtel PC's with Mandrake Linux pre-loaded been? Do you think that Microtel's actions might lead additional second or perhaps even first tier OEMs to take similar actions?

GD: There has been very positive feedback about Mandrake/Microtel PCs. They are reasonably priced, powerful and Mandrake has been well integrated in the machines.

On the other hand, HP/Compaq also offers Mandrake Linux as an option for several of their workstations. Linux on the desktop is really starting to move quickly, with a lot of demand coming from corporations who wish to migrate their workstations to Linux. Several other deals are also in the pipeline. But Linux/OEM remains a big challenge because there is a lot of pressure from Microsoft to prevent its success.

We also ship Mandrake in an Advantech appliance called "Advantech Firewall Plus". It's a great router/firewall powered by a special security-focused version of Mandrake. But for now the product is only available in a few European countries.

OfB: In a recent Linux and Main interview [ http://www.linuxandmain.com/modules.php?name=News&file=article&sid=141 ], "Rasterman," of Enlightenment window manager fame, is quoted as saying the Linux desktop is dead, and that the future of GNU/Linux is in the embedded space. What are your thoughts on this statement?

GD: I remember quickly reading that interview. I think many companies that produce embedded systems will benefit from using Linux as a software base component, because Linux is free, modular and easy to hack. But I have doubts about creating a profitable company based solely on producing Linux software for embedded systems, unless the product was released as proprietary software.

About the death of the Linux desktop, I think Rasterman is wrong. What I see currently is the birth of Linux in the desktop area.

OfB: When you started out Mandrake's biggest feature was the addition of KDE to the Red Hat base distribution. These days Mandrake seems more neutral between the KDE and GNOME desktops; do you still see Mandrake as being KDE-centric?

GD: Mandrake is definitely not KDE-centric anymore, nor is the distribution based on Red Hat -- this has been the case for a very long time. In 1999 I was convinced by Jacques Le Marois - our CEO - to offer GNOME and other graphical interfaces such as IceWM and WindowMaker with Mandrake. It was a great idea, and this lead me to later write an essay about "Diversity is the strength of Free Software".

Our brains are conditioned by practice, especially proprietary software practices. As a result, many people come to us and ask "Why offer several graphical desktops - it's silly!". But after a while they come back and tell us that even though they personally prefer KDE, their spouse or children prefer GNOME, etc. My opinion is that by offering many options we are answering the largest number of various needs and preferences, so it's good for everyone.

OfB: Do you see Mandrake offering more server related tools in the future? Some people have noted that Mandrake Control Center has less server related functionality than other tools such as SuSE's YaST; do you see that changing?

GD: The Mandrake Control Center is a very modular tool. Depending on which installation class is selected, you will have more or less configuration options. For instance, if you choose to install a workstation, you won't get all the server options. Additionally, we try to improve it all the time!

OfB: Any closing thoughts you'd like to leave us with?

GD: I've very happy you asked me so many interesting questions, it's not so common. Thank you very much.

OfB: Thank-you for your time GaŽl.

Copyright 2002