Interview with Raphaël Hertzog
Today I’m presenting you my first interview, today is the turn of Raphaël, a Debian Contributor.
I’ve meet Raphaël on his blog. He is interviewing other Debian contributors, and I’m interviewing him, continue reading and know him better.
Who are you
Hi. I’m 31 years old, I live in France near Saint-Étienne. I’m married and have a one year old son.
I joined Debian in 1998 while I started my studies at the engineering school (INSA Lyon). I graduated in 2002 and started working for a Swiss company where I created a custom embedded Linux distribution (based on debian-installer) and developed customer-specific proprietary software for use on its products.
We’re now in 2010, I’m preparing the fifth edition of my book (for Squeeze) and I’m still doing free software consulting/development with my own company.
What’s your biggest achievement within Debian or Ubuntu?
Reviewing my history to decide what to mention…
In terms of number of hours put in a single project, I think it’s the development of the new source formats and all the associated work to get them accepted in the archive, and to smooth the migration path for all maintainers.
In terms of impact on Debian, my proposal of “sponsoring” package uploads evolved and ended up really changing the way we deal with prospective maintainers. That was 11 years ago and we still have sponsors to review and integrate the work of newcomers.
What are your plans for Debian Wheezy?
That’s difficult, I have always more ideas than time to implement them and the set of ideas that are likely to be implemented tends to vary over time. :-) So take my plans with some care…
Let’s start with dpkg. I would like to provide a standardized way to manage a Debian source package in a VCS with the goal that the VCS contains all the information required to generate a proper “3.0 (quilt)” source package (including upstream patches with DEP-3 headers). I wonder also if we shouldn’t switch dpkg-deb to use XZ compression by default, we could save a few gigabytes on the mirrors and put more software on a single CD. I also want to improve dpkg-buildflags so that it can be used to inject hardening build flags (cf #489771).
I also have plans for bigger changes concerning Debian, and among them is the introduction of Debian Rolling, a distribution similar to testing but with some design choices to make it more usable at any point in time.
If you could spend all your time on Debian what would you work for?
I would try to fix the 400 bugs filed against dpkg.
I also have a quality assurance project: I think we have to detect sooner when a package is no longer actively maintained, i.e. way before users start to complain about the state of the package. We have many ways to evaluate packages, we should combine them in a sort of “maintenance mark” and have a semi-automated process to ping maintainers when packages drop below a specific limit. The collected information can then be used to find new/supplementary maintainers.
Another pet project of mine that keeps coming back in my head is what I call “Debian Maintainer Information Hub”. The goal would be to have a single service collecting all the information that a maintainer is supposed to get. Every contributor can then decide what information he wants to receive and how he gets it (mail, batched mails, RSS, …). To achieve this, many services should thus stop sending mails to the maintainers and instead send them to this new service. It would require work on many parts of the actual infrastructure.
What’s the biggest problem of Debian?
Our project identity is somewhat minimalistic. It evolves around the social contract and the Debian Free Software Guidelines. Both documents answer the question of what we’re doing but we lack a clear answer to the question of how we’re supposed to work towards our goals. It would be great if Debian could agree on some principles concerning topics like goal setting, collaboration, team work, politeness, respect. We could then advertise those and build on them while recruiting volunteers.
What’s the biggest problem of Linux?
I’m always annoyed when some new hardware doesn’t work. It’s painful to have to google around to find out a solution. Sometimes you need a newer kernel, sometimes you need to build custom modules for your kernel, sometimes you need a firmware for a driver that you already have, sometimes you need a newer version of Xorg, etc.
I wish all hardware manufacturers could get their drivers merged in Linux before their hardware hit the shelves of shops.
Do you think the success of Ubuntu is beneficial for Debian or the opposite?
I think it’s mostly beneficial. There have been many challenges on the path, but nowadays collaboration is encouraged from both sides and the ties between both projects are getting stronger and stronger due to the numerous Ubuntu contributors who are joining Debian (both as maintainer and as developer), and vice-versa.
But Debian should definitely try harder to harness Ubuntu’s popularity to get more attention on itself.
You might want to look at a recent LWN article on the topic: State of the Debian-Ubuntu relationship.
Beside Debian and Ubuntu which other distributions you like or admire? Why?
I must say that I am somewhat exclusive in my relationship with Linux distributions. I have not tried any other distribution besides Ubuntu since I joined Debian. So it’s difficult for me to say that I like or admire another distribution.
I follow Fedora’s development through the regular updates I get on Linux Weekly News and I like how their bleeding-edge focus contributes to iron out rough edges very early.
Which is your favorite Desktop or Window Manager? Why?
I use GNOME with the Awesome window manager. I switched from metacity to awesome because I like to use all the space on my screen. Before that I was using terminator to have multiple terminals in a maximized window but it was only solving the problem for one application.
Why GNOME? Because it’s the default desktop in Debian and I’m just used to it.
What other projects in the Free software community are you working on?
But if I had time, I have a few ideas of software that I’d like to write. One of them is named Librement and is a sort of assistant to help you contribute to free software.
In the past, I wrote Geniustrader but I’m no longer an active stock trader so I handed over the maintenance.
Anything else you may want to add?
Yes. I decided a few months ago to give a new direction to my career: I want to focus on my writing activities. That’s why I gave a new life to my blog, you’ll find at least two new articles each week. I provide useful documentation for users and for contributors (prospective or not). That’s also why I’m going to translate my best-seller Debian book into English.
I invite you to subscribe to my free newsletter, you’ll get my monthly summary of the Debian/Ubuntu news, and you’ll be informed of the progress of my book projects.