Branden Robinson, outgoing Debian Project Leader, was kind enough to answer
a few questions via email.
LWN: Now that your term is winding down, do you feel that you have
accomplished what you hoped to accomplish?
No. In and of itself that is not a bad thing; it's better to have a
surfeit of ideas than a paucity of them, but even so I found the position
to be a subtly different kind of challenge than I expected.
Still, I learned a great deal about the inner workings of the Debian
Project's infrastructure that I don't think I could really have come to
understand any other way. I look forward to being a resource for the next
Debian Project Leader.
The important work that Debian has to do will continue.
Project Scud was announced almost exactly one year ago. Since then, has
this project helped with the management of the Debian project?
In ways, it has. The DPL team was valuable to me in my role in that I
think it was essential in either keeping me informed about various
behind-the-scenes happenings, or in offering differing perspectives on
things I already knew about.
Should some variant of it be continued into the coming year?
I think so, yes -- however, the DPL Team had a big problem with visibility
to the Debian Project at large, and that was a significant liability.
It felt to me like the DPL Team was constantly engaged with somewhat
sensitive personnel issues that were difficult to air publicly in a way
that was both constructive and fair to all the parties involved.
The level of harmony within the team, however, was very high, and it was a
good working environment. We all exhibited respect for each other and were
able to work constructively even where we had differences of opinion. I
had been afraid that we wouldn't gel, and with the exception of one member
who simply didn't (and doesn't) have the time to participate, I think we
Quite apart from who wins the DPL election, I value the stronger
relationships I've forged with Debian developers over the past year, both
within and apart from the DPL Team.
When it comes to the team approach being continued into the coming year, I
think it's inevitable. Whether it's called "the DPL team" or "Project
Scud" or doesn't have a name at all, but is instead "the guys the DPL
drinks with at the pub", I feel certain the concept will continue to exist
in some form, just as it predates its explicit identification last year.
The role of DPL is a multifaceted one, and it's just plain good leadership
to share the responsibilities. Just as the DPL has the trust of the
developers, so too must a DPL demonstrate trust in others. The best
leaders find ways to trust new people, rather limiting their horizons.
Are there things you are particularly happy about? Or particularly unhappy
I'm particularly happy that the day-to-day machinery of Debian, of package
maintenance, quality assurance, release management, propagation of
unstable packages to testing, and so on, continued to hum along as it
should. Debian's technical processes are, for the most part, highly
developed and mature, and not something the Project Leader needs to meddle
with. That was deliberate in the design of the Debian Constitution, and I
think that is a point of continuing success.
The Sarge release, and, critically, the maturity of the d-i
(debian-installer) project are also achievements I'm enthusiastic about. I
don't claim credit for them in my capacity as DPL, except insofar as I was
smart enough to know not to meddle with something that was working. Our
release management processes have started to seriously hum over the past
year. I think we really have a handle on management of major transitions.
The BTS has seen major improvements, the devscripts package has more useful
tools, and more people are leveraging these new features to get their work
On the downside, I'm particularly unhappy that a few particularly thorny
issues occupied virtually 100% of my time. I made a conscious decision
even before I was elected to grapple with what the Project has identified
as the most critical issues, not necessarily those where I could make a big
splash for myself or grab headlines.
One consequence is that things I have achieved are difficult to measure;
another is that I didn't have much time left over to work on even the
somewhat strange things I consider "fun", like coming up with a new set of
trademark usage guidelines. That's still being managed ad hoc, and it
doesn't really need to be.
It pays to keep in mind, though, that the most visible thing Debian does is
get free software to our users. That's the primary mission, and every time
I dwell on my frustrations, I need to remind myself that Debian is
fundamentally succeeding in that mission. The free software landscape is
littered with the remains of projects that have failed in it.
Consequently, it is invaluable to maintain one's sense of perspective.
Why did you chose not to run for a second term?
There are factors on a few fronts. As you may gather from my previous
answers, I have a bit of battle fatigue. More importantly, however, I have
come to appreciate the wisdom that a few people in the Debian Project have
already expressed. First, you don't necessarily have to be the DPL to get
things accomplished. The DPL is not a strong executive under our
constitution, and some of the DPL's constitutional powers, such as the
dismissal of a delegate against his or her will, have never been exercised.
Secondly, many developers don't seem to really appreciate the first point.
It's often been remarked that the Debian Project only seems to seriously
grapple with internal management issues once a year during the elections.
In between, most people seem to just wait for the Project Leader to pull a
rabbit out of a hat.
While it's certainly possible that a more talented leader than myself could
do so (or simply be the straw the breaks the camel's back), it would be
healthier if more developers were more involved with those issues.
What I'd like to do next is see if I can mold myself into an example of
what I'm beginning to think of as the "good Debian citizen". I've had the
benefit of an "insider's view" of what's right and wrong at the core of the
Project -- what I think is critical now is to better uphold clause four of
our Social Contract, in which we commit to openness with our users. That
clause talks specifically about bug tracking, but many within the Project
think we should apply it more generally.
Some Debian developers have an ambivalent relationship with the Project's
"insiders" because, simultaneously, they are details of infrastructure
management that most of them don't care to know about -- except when
they're perceived as not working. In that case, they demand satisfaction.
I don't particularly decry this so much as note it to be human nature.
What do you think will be the biggest challenge for the next DPL?
Infrastructure reform, which seems to eat every DPL that dares to grapple
with it, will threaten to do so with the next leader as well.
Is there anything else you would like to add?
It has been a tremendous honor to serve my fellow Debian developers and
users in this office. I've had a few opportunities to speak before
audiences familiar with Debian during my term -- the Open Source World
Conference in Màlaga, Spain, and at Free Software and Open Source Days in
Istanbul, Turkey, are two recent examples.
Everywhere I go in my capacity as a Debian representative, I meet many
people who have boundless enthusiasm for the Debian Project and the work
that we do. In many cases these are people who are as young as I was when
I started using Debian, ten years ago, or even younger. Many of them want
to be involved but want advice on how to contribute -- they don't know if
they have anything to offer the project. The advice I offer is simple:
identify something you care about, where your natural interests tend to
flow, and throw yourself into it. A GNU/Linux distribution is an
infinitely improvable thing -- that is, we're never going to run out of
ways to improve it. When there aren't features to be added or bugs to be
fixed, there are translations to be made, documentation to be written, or
licenses to be fixed. It seems basic to Debian old-timers, but it's a new
insight to Debian's vigorous youth.
At the GPLv3 launch conference in Boston this past January, I troubled Eben
Moglen for a recipe on how to grow the Free Software community. His advice
was simple, as most good advice is: "Each one, teach one." Over the past
year I've been able to impart just a little bit of my meager knowledge to a
great many people. That has been the most rewarding part of this job.
to post comments)