Q&A: Clement Lefebvre, the man behind Linux Mint

The creator of the world's most popular Linux distro talks candidly about his goals, his successes, and his nightmares

Linux Mint, an Ubuntu and Debian-based Linux distribution, has seen tremendous growth in community support and installed base in recent years. Since arriving on the scene in 2006 with its first release called "Ada," Mint has become the most popular FOSS operating system on DistroWatch.com, surpassing both Ubuntu and Debian themselves.

Mint is available with out-of-the-box multimedia support and now even has its own desktop interface, Cinnamon. Freelance writer Christopher von Eitzen interviewed Project Founder and Lead Developer Clement Lefebvre about Mint's origins, major changes to the distribution, its growth and its future.

[ InfoWorld presents the Bossies 2013, the best open source software for data centers, clouds, mobile, and more. | Get the latest insight on the tech news that matters from InfoWorld's Tech Watch blog. ]

[TEST:Linux Mint 15 delivers smooth alternative to Ubuntu]

What is your professional background and what was the first Linux distribution that you ever used?

I got a Masters in Computer Sciences from the University of South Paris in 2001. I was mostly interested in game development, but as it happened, I worked for banks, telecom and software companies in France and in Ireland. I had various job titles (web developer, IT engineer, software developer, J2EE architect), and in one company I was teaching rather than coding, but most of time my job was to design and to develop software or web applications.

My first distribution was Slackware. In 1997, if I remember well, a student at my university brought a shiny set of Wallnut Creek floppies. Everybody got excited at the idea of having a Unix system at home (we were developing on IBM AIX at the university). I got immediately hooked. Slackware was (and still is) a piece of art, clean, predictable... and it was also my first encounter with Free Software.

Very few people ran Linux back then. You had to own a "Sound Blaster" card for audio to work and it'd take a novice a week or two to get the mouse to work and achieve a good resolution with X11. And of course, there were very few applications available. Most users were university students, teachers, or developers with a serious taste for adventure and a strong immunity to discomfort. This, and the novelty of the Free Software ideas were very enlightening to me and I really enjoyed being part of this in the late 90s.

You launched the first release of Linux Mint, code-named "Ada", in 2006. What made you want to fork Ubuntu and create a new distribution? Why Ubuntu and why the name "Linux Mint"?

I was writing for http://linuxforums.org and I wanted to try and host some of my articles myself so I needed a domain name and I chose "linuxmint.com" (mostly because it was short, obviously related to Linux, and connoted the notion of freshness and technology which just works). That's how the name came initially, as a domain name for a Linux website.

Parallel to that, I had been reviewing distributions for a while and tinkering with a lot of things so I had good knowledge of how they worked and what I liked or didn't really like in each one of them. At some stage I got something out and I wrote a bit about it on linuxmint.com. I was surprised by the feedback I got and a few months later it was clear people were more interested in that little project than in my articles.

There was no plan to create a new distribution initially, and certainly not one that could rival the likes of Microsoft Windows, Apple Mac OS or even the established Linux distributions. People showed interest in what I was doing and so I responded to that.

Ubuntu was never "forked". It was and still is used as a package base and regarded as an upstream component. Why Ubuntu? Because it was (and still is) the best package base.

What were the original goals that you set for the initial version of Linux Mint and did you achieve them all? And what would you have changed about it if you were given the chance to do so?

The first version had no goals at all. It was something I experimented with. After that, feedback shaped versions 2.0, 2.1 and 2.2. It's only after Linux Mint 2.2 "Bianca" that things got serious. 

The biggest goal for me is to make sure each release is the best release to date. It might sound funny but it's not always easy to achieve. Things move fast upstream, technologies change. My main concern is to continue to provide to people what they like and to get closer to a better desktop, to make it easier and more pleasant to use, each time.

What makes Linux Mint different compared to Ubuntu or Debian beyond the visual changes?

I'm not really sure. It never "had" to be different. We're already radically different in the way we work, in what we think is or isn't important, in the vision of the desktop we have, in how we look at security and package updates... so with all that we've often evolved in opposite directions. Now, from a technical point of view, if you look at Linux Mint, it is either using an Ubuntu or a Debian package base, so beyond the desktop layer (where we're primarily focused) you'll find very few differences.

To a certain extent, and despite a fork of their package base, it's also the case between Ubuntu and Debian. We're talking about the same base but the two distributions are very different, primarily because they're focused on very different things. Ubuntu patches the desktop and user applications layers heavily and at times they introduce significant differences in the core of the OS (locales, Upstart for instance...).

Both Ubuntu and Linux Mint evolve according to their own vision and consider upstream components as ways to achieve that vision, which means things don't have to be different but changes can be introduced when needed. In the case of Linux Mint, we try not to introduce incompatibilities with upstream package bases, so we keep the same libraries, same versions and we do not fork the base.

In recent years, Linux Mint -- currently ranked higher on Distrowatch.com than the distribution it is based on -- has steadily increased in popularity among both beginners and expert Linux users. How has this constant influx of new users changed the direction of the project, if at all, and what problems have you encountered because of its quick user-base growth?

The biggest issue is to scale. Some aspects such as hosting are easy to solve (you just throw more money and resources at it), others such as the quality of the communication between the team and the community are much more problematic. Thankfully Linux Mint has always been growing, so this isn't new. We might be X times the project we were back in 2006, but we've always had to scale so we learnt a lot and we continue to learn from it as we grow larger.

This year we introduced a new blog and we recently restructured the way we host packages and deliver Debian update packs.

What was it like transitioning from GNOME, used in Linux Mint 2.0 to version 12, to the Cinnamon and MATE desktops with the release of Linux Mint 13?

It was a nightmare but it's also something we're very proud of.

Some people don't understand how important GNOME was for us. It was by far the most popular and the most mature desktop environment. It was a huge component for us. Most of the technology we had developed was built around GNOME and as a desktop distribution we had improved incrementally around it since 2006. So there was simply no way we were going to stop using it.

We have a vision and no matter what happens we get things done. This was the biggest challenge we faced since 2006. Linux Mint 12 was hugely impacted and it shipped with an early version of MATE and a set of GNOME Shell extensions which roughly did the job but weren't any long term solution. This was really tough for me personally. You can imagine what happened during that development cycle... all efforts, all resources, everything we had went into getting back on our feet. When Linux Mint 13 "Maya" was out, I had a huge smile on my face, MATE was stable and Cinnamon was there. I'll probably always have fond memories of Linux Mint 13 as it marks the end of the drama and the recovery from the loss of GNOME for us. "Maya" is also my daughter's name, so that adds to it even more. :)

With the latest stable release -- Linux Mint 15 "Olivia" -- out the door at the end of May, what changes and features are you most excited about?

Linux Mint 15 brought a lot of features people were asking for; a more interactive login screen (MDM HTML5 engine), unified settings in Cinnamon, an independent Driver Manager, proper repository and PPA management...etc. For the first time also, Cinnamon handles the entire visual layer of the desktop on its own including the control center and the screensaver, no part of GNOME 3 is visible.

What would you like to see make it into version 16 "Petra", which will be based on Ubuntu 13.10 to be released in October?

The Software Manager will feature performance and speed improvements. We'll introduce a USB stick formatting tool. Cinnamon won't be a frontend to GNOME anymore but a complete desktop environment with its own backend. Cinnamon 2.0 will also introduce improved window tiling, a new feature called window snapping, and better user and session management. These are the features which are ready for Linux Mint 16 at the moment. There's also a roadmap and a lot of further improvements planned for the release but I'd rather not talk about these until I'm sure they'll be implemented.

While Ubuntu offers an automated in-place upgrade option to install the latest major releases of the distribution, Linux Mint does not officially support this functionality. Instead users are advised to backup their data, perform a fresh install of the current stable version and then restore their data. Why has the team decided to forgo this option and are there any plans to include official support for in-place upgrades in the future?

Well, it's possible to upgrade Mint the same way you upgrade any Debian-based distribution, including Ubuntu, and I'm sure it works pretty well. That's not to say it's a good idea to do so. First, few people are experienced enough to troubleshoot problems related to APT. Second, it takes more time and bandwidth to perform an APT upgrade than to download and install a new release (which 900MB ISO can contain between 3 and 5GB of compressed data). Third, when you install from a live system you get a unique opportunity to see the new release, to test the new kernel with your hardware and to make sure things work fine before you make the jump. Now with this said, things can certainly be improved. We should probably insist on people creating a /home partition during the installation, we should probably implement safeguards on UID and permission checks after a fresh upgrade... there's definitely work to be done for upgrading to be made easier. Ubuntu's recommended solution isn't something we want to back though, it's not good enough for us to recommend. Automation is one thing and making a process trivial is usually an improvement, but when that process is risky, automation is really dangerous.

With Canonical pushing for its Mir display server technology and XMir, the X server for Mir, for Ubuntu 13.10 and beyond, what does the Linux Mint project plan to do? Will it follow Ubuntu's example or switch to Wayland? 

Whether it's in the scope of Linux Mint or Cinnamon, we're only interested in stable and proven technology. If tomorrow Mir or Wayland arrive in our repositories and we can use them to give people a better experience without causing significant regressions, we'll consider using them. Ubuntu 13.10 is pushing for Mir so we'll probably test it and see if it fits, how mature it is and what are the pros and cons involved in making use of it.

In September 2010, the project released the first Debian-based edition of Linux Mint (LMDE), replacing the distribution's standard Ubuntu base with Debian's testing branch. Why create a Debian edition? How popular has it been and what percentage of Linux Mint users are running it? What are the project's future plans for LMDE?

We wanted to know more about our package base, as a component. So the best way to know more about it was to switch it to another one and compare the differences. What would we lose if we stopped using Ubuntu? What did it bring to us compared to vanilla Debian? It was important for us, as a project which innovates and develops a lot on the desktop layer, to see how portable our technology was. Could we port it to other bases? Would it work out of the box elsewhere? Or was it too tightly tied to Ubuntu?

We wanted to know more about the pros and cons of frozen release cycles vs. rolling models. We learnt a lot from it, so from an R&D perspective it was a huge success. As a Linux Mint edition, it's also relatively popular. If we gather all LMDE users, it comes in third behind the Cinnamon and MATE editions. Our plan is to innovate on the frozen cycle and to develop continuously for Linux Mint n+1. LMDE follows a semi-rolling path with update packs and benefits from all the development done on the latest Linux Mint releases. The goal for LMDE is to continue to be as similar to the mainstream editions as possible and to feature the same improvements release after release with Debian and without Ubuntu.

Do you ever foresee dropping Ubuntu completely in favor of Debian or another distribution for the basis of Linux Mint?

1 2 Page 1
Page 1 of 2
InfoWorld Technology of the Year Awards 2023. Now open for entries!