April 20, 2005

Interview: Branden Robinson, new Debian Project Leader

Author: JT Smith

On April 17, Branden Robinson, a senior employee at
Progeny Linux Systems,
took over the role of Debian
Project Leader (DPL) from Martin Michlmayr.
For Robinson, the moment was a long time coming.
Having run unsuccessfully for DPL every year since 2001, Robinson
had taken to making joking comparisons of himself with William
Jennings Bryan, the perennial candidate for the American presidency
in the first decades of the 20th Century.

Beneath the jokes, Robinson had become increasingly pessimistic. This year,
he had seriously considered not running. He expressed his uncertainty on a
web page,
and stated that he wouldn't run unless 100 Debian Developers wrote
comments signed with their GPG keys supporting his candidacy.
Whatever Robinson's reasons for this request, it
seems to have gotten votes. Even
controversy about Project SCUD,
a self-appointed group of advisors to
the Debian Project Leader to which Robinson belongs, was not
enough to prevent his election. When results were tallied according to
the Condorcet method
on April 10th, Robinson had won by a narrow margin over six other
candidates, with just over 52% of eligible Debian Developers voting.

Robinson has been one of the most active Debian
Developers for at least five years. He's been involved with the project since 1998 and now maintains over 125
packages, most of them connected to XFree86. Since 2001, he has been
on the Board of Directors for
Software in the Public Interest, the non-
profit organization that supports Debian. In the past, he has also been
active on the debian-legal list
and Debian IRC channels. While Robinson's work is respected by many in the
Debian project, he has also been accused of being a flamer, sometimes unjustly,
as when his obscure reference to
The Great Santini
was interpreted as a deliberate insult to
Debian users. At times, too, he admits, his IRC handle "Overfiend"
-- a reference to an infamously violent, misogynistic
anime film
and its sequels -- probably attracted unnecessary controversy.

In person, Robinson is a small, overweight man just over thirty. He
comes across as astute and earnest. In fact, he has the general air of
many social or political activists. A sharp, often self-deprecating sense
of humor is obvious. Over the phone, he sounds more articulate and
noticeably more self-confident than when I first met him five years ago.

Recently, Newsforge talked about his election, Debian's relationship with the
business community, the problems he sees in Debian, and what he hopes to
achieve in his new role.


The Election Campaign

Newsforge: Congratulations on the election. But one thing I have to ask:
why did you keep running when you'd been defeated so many times?

Branden Robinson: I guess one answer is obstinacy. Persistence would be
another.

However, essentially, the reason is that the problems have not
changed. It's something of a truism that the most difficult thing for
businesses to do is manage growth successfully, and that truism
also applies to the Debian project. The number of architectures that we
support, the number of source packages in our distribution, the number
of developers that we have -- all of these have contributed to difficulties in
simply getting our work flow smooth and keeping it smooth.

In general, there has been a lack of progress within Debian for several years.
The project has had a series of leaders that, for one reason or another,
not necessarily through any fault of their own, have been unable to make
much progress in improving the situation. In my view, that is cause for frustration.

NF: Has Debian made any progress at all since you started running?

BR:The New Maintainer system [the process for integrating new
Developers] is a little creaky in some people's view, but I think a big improvement
over what we had before.

NF: Have your priorities changed since you started running?

BR: I used to see myself as a kind of iconoclastic reformer who was going to move in
and kick ass until things got fixed. That provoked no small amount of resistance.
As I've gotten to know people better and gotten
more insight into the infrastructures of Debian, I have realized that's a bad approach.
It implicitly impugns the motivations of the people doing the work. My goal
now is bring problems to light and get what I can to help those teams scale
to the challenges that face them.

NF: Looking over the past elections, I notice that your public presentation of yourself in
e-mail and on IRC has changed considerably.

BR: I can still talk to peel paint on IRC when the mood strikes me. But in large part it's a matter
of knowing your audience. The shift that people perceive is in What Is Branden Like When He's
Not Running for DPL. I've changed because I needed to change.

NF: In the past, you had to live down your IRC nickname of Overfiend. Was that a problem in the
2005 election?

BR: The issue that I took a pasting over this year didn't have anything to
to do with writing flames or having an outrageous IRC nickname. It was at least more relevant.

NF: That would be Project SCUD. I know that Debian members were concerned that, if
a DPL was elected who supported Project SCUD, then an unelected cabal would be making decisions.
This view may have aggravated by the fact that both two candidates, you and Andreas Schuldei,
were members of Project SCUD, as well as former DPL Bdale Garbee. To some, it apparently
seemed as though a palace revolution was underway.

Others were concerned that nothing in the Debian constitution authorized such a group, and
whether membership in the group would be closed. What are your thoughts about Project SCUD
and the controversy it raised?

BR: I described
my view of Project Scud
a bit during the campaign, but to flesh it out further:

  • It's there to help the elected DPL not lose track of important things
    that need to get done.
  • It's there to help provide insight to the other members into areas of the
    Project they're not personally familiar with. It's not just
    the DPL who benefits.
  • It's there to serve as a sounding board for the DPL's brainstorms in a
    way that won't cause mass panic if he puts forth some harebrained idea.
    I'm conscious that my ruminations, even if I'm not doing them with my
    DPL hat on, may be misperceived as initiatives I'm bent on accomplishing.
    There will less time wasted on flaming and hysterics on the lists if someone
    like Bdale Garbee or Steve Langasek [another Project SCUD member] can say
    to me privately, "uhmmm...that doesn't sound too clever".
  • It's there to serve as a source of personnel to serve as representatives
    of the DPL in the event the actual DPL isn't well suited to negotiate
    with a certain party. I do have a certain history that isn't
    entirely positive. While I'll do my best to be diplomatic and
    try to overcome this history, sometimes the most productive
    tactic is simply to rely on someone else to represent me.

This sort of group has always existed anyway. I mean, I doubt
that Martin Michlmayr stopped talking with other Debian people
at Cambridge down at the pub just because he became DPL.

And while I do believe my election to be an
expression of a desire for a more hands-on, more visible, and more
assertive project leader than we had with Martin Michlmayr,
I don't interpret it as some sort of mandate to get into
people's faces, yell and scream, publicly (or privately) humiliate them,
and bring them to heel.

I do have an agenda to serve, and that's stated in my
platforms.
You'll note that "extract postures of prostrate subordination from people
who've argued with me in the past" is not among my goals.

I want to achieve my goals while making the Debian Project a happier place
for everyone in it. That means being conciliatory, polite, and friendly
wherever I can be, and having something resembling the old "Overfiend rage"
persona only in reserve as a last resort. I'm hoping to be able to focus my
obstinacy and tenacity in good ways, not bad.


Debian and Business

NF: In your platform statement, you mention that you have some managerial experience
in your daily work at Progeny. Can you elaborate?

BR: My role is as team leader of the Platform Services group. People don't
really report to me for sick days. But I have been responsible for keeping track of people's time,
and assigning them tasks and knowing when to delegate something and when to handle it
myself, and having responsibility for delivery.

I've also been involved in the interview process and candidate selection for new hires, and
I'm consistently involved in evaluation of new prospects for
customers -- doing technical evaluations, strategic evaluations in some measure.

NF: How will this experience in dealing with Debian commercially help you as DPL?

BR: It gives me an important perspective that I didn't have a few years ago. A lot of Debian
consumers are, in fact, companies. That's something that might have seemed strange to
me some years ago when I first joined. Then it seemed a distribution for geeks by geeks.

Over the past few years, my thinking has changed. The market thinking has certainly changed:
Debian is not too alien, and not too strange -- it's a product that in some senses can be a service. For
example, if you look at Debian unstable, there's this ongoing wellspring of fresh packages. You can think of
it almost as a subscription service that you don't have to pay for. That is something that makes sense to a lot of
business interests.

NF: Do you think your perspective about business is unusual in the Debian project?

BR: No, but there might have been a time when when I would have
thought so. I think integration with the business world is something that is pretty obvious.
There's always been a large contingent of business-friendly developers in Debian. It's not
all a bunch of communists, and I don't think it ever was.

NF: Still, I suspect that many people think of Debian as the ultimate non-business distribution.

BR: The culture of Debian tends to be a little more suspicious of the compromises that the
business environment asks people to make, rather than being opposed to business per se.
There are certain distribution models that some people feel are immoral or unwise
or both, but it's never been about taking down the corporate colossus for its own sake.

NF: A common perception in the free software movement is that some FOSS companies are
profiting from other people's work without making any contributions to the community.

BR: I think there's a form of that type of resentment in the free software community, but it's
not the same that you see in the traditional copyright market, for instance. Take the file-trading
networks. The large copyright interests are keenly intrigued by the prospect of a pay-per-experience model --
every time you listen to a song, you pay a fee. At the very least, they want to catch you when you download it.
In the free software community, nobody cares to track that kind of thing.

It's very much a matter of how you want to look at things. Large companies have an economic interest
in looking at things one way, while free software developers have a pretty
different way of looking at it.


Issues Within Debian

NF: In March, Steve Langasek released a summary of several days of discussion by archive and build
administrators and the release management team. You've called this summary the
Vancouver Prospectus.
The summary included suggestions for streamlining the next Debian release by concentrating on key architectures
and de-emphasizing others. These suggestions have generated hundreds of emails on the Debian lists. What's
your opinion of the Vancouver Prospectus?

BR: I mostly think it's controversial. I thought that even before it was posted publicly. I knew it would make
waves because of what it concretely proposed, and because of the way that it was presented, which was
fairly terse and probably contributed to the perception that it was a done deal. But as I said on the mailing list, this
is the beginning of a process to figure out where we're going to, not the end of it.

NF: So you see it as a basis for discussion, not a final policy.

BR: That's correct. If some of the people concerned with those architectures that are demoted to second class
(not very diplomatic language for this sort of thing, but we'll go with that) move forward and do what is asked
of them, and meet the requirements for the architectures that are remaining in the system, they'll be restored
to first-class status.

NF: So basically the level of interest and activity will be determining each architecture's status.

BR: Yes. I don't think there's any ultimatums here. And nothing's being asked that's terribly
unreasonable. For example, there don't seem to be any active alpha supporters right now. The alpha architecture is ported, but nobody is actually serving as chief point of contact. And that's a big challenge to maintain an architecture when you don't have anywhere to go with architecture-porting issues.

NF: Another issue raised in the
leadership debate
was the need to encourage diversity within Debian. One aspect of this problem is
being addressed by the Debian Women group, but some people are suggesting that more effort could be made to
accommodate other languages as well. What are your thoughts on the question of diversity?

BR: Being of a liberal bent, I like the concept of diversity pretty much on its own merits. Besides which, in any
attempt to diversify Debian, we may draw in people who are interested in working on usability issues.

NF: The evidence is inconclusive, but I've heard it suggested that women in free software are more interested
in education and outreach than the men are.

BR: I wonder, though, how much of that is simply a consequence of self-identifying as a minority group? If you are
a minority group, it's almost inevitable that you'll be putting some sort of emphasis on education and outreach.
You're trying to address that minority status, either by raising people's awareness about it --which is education -- or
by making yourself less of a minority -- that's outreach. That said, I'll go out on a limb and say that the Debian Women
group will have a large and unique impact, and probably a positive one.

NF: During the debate, you suggested that diversity could be encouraged by making the Debian project a less
threatening place. How would you go about doing that?

BR: A lot of the problem is that Debian's growing too large. There's probably not a single developer who's
met every other developer, and that's been true for a long time. It was interesting that after the election I got
congratulated by people I had never heard of before. That was part of the humbling of the experience.
It reminded me just how broad my responsibility for representation is. I'm answerable not just to the
people I know from the mailing lists, IRC channels or conferences. I'm accountable for and responsible for
a number of people, and I'm not always entirely clear who they are. They may be Debian developers, but they
may also be Debian users. Either way, I have to serve their interests.

In terms of a threatening environment, I think it's mainly a matter of playing the ombudsman. I don't want to go out
and lay down the law in terms of making a blanket statement such as forbidding the use of Debian in the name of
an IRC channel without prior authorization or anything. I mean:

a.) That would be heavy-handed and probably make things even worse, and

b.) That would create the illusion that it could actually be managed.

But in particular cases there are probably steps that can be taken to mitigate the influence of really hostile people.
When those are brought to my attention as leader, I'll act on them. But I'm nervous about too much pro-activity,
because it could worsen the problem by giving it higher visibility.


First Steps and Measuring the Results

NF: What's the first thing you want to do as Debian Leader?

BR: Not to get in the way of the Sarge release [the next version of Debian] -- but that's kind of a non-step. My first step
is simply to go around and introduce myself to a lot of the groups and talk to them. I expect to
be doing a lot more listening than talking initially, especially internally.

I'm interested in finding out what concerns the various internal management groups that
already exist have. In other words, what are their priorities, their areas of concern? What would they
like me to do? I also want to reemphasize that I'm not here to perform a Stalinistic purge. I'm here
to find out how I can make Debian better by asking people who should have some ideas on this front.

NF: How long will this stage last?

BR: I'd like to have a fair amount of feedback in time for my first monthly report, around May 17. I'd
like to spend a few weeks just getting oriented. I've spoken to Bdale Garbee and Martin Michlmayr, and
there's also a transition that takes place between the last DPL and the next. There's things that
I may not have been aware of that have been being handled quietly.

NF: What would you like to do as DPL?

BR: Historically, there's been a lot of complaining about the roles of the Debian
account managers, archive administrators and system administrators. I want to be in the position
to explain and defend how and why things are done inside Debian.

NF: Will you be representing Debian publicly?

BR: Progeny has given me some flexibility to attend more trade shows than I have been.
That's something I'm looking forward to doing. I'm accustomed to public speaking, I'm comfortable with it,
and I like to meet new people and hear what they have to say. I hope that I can persuade them that Debian
is a more exciting and vital project than they thought in the first place.

NF:Will you make a point of visiting certain areas of the world to help with the diversity issue?

BR: No, I'm going to let the invitations come to me and achieve diversity through what I'm asked to do.

NF: Do you have any plans for closer ties with Debian-derived distributions, such as Knoppix, Ubuntu and
MEPIS?

BR: We probably don't have as close a relation with spin off distributions as we should. I'd like to see
what we could do to strengthen those bonds a bit.

NF: How will you measure the success of your efforts?

BR: One measure is the old aphorism about leaving it better than you found it. I think having Sarge
released would be a measure of success, although not one for which I could claim a huge proportion
of responsibility. But if I have de-mystified some of the internals of Debian, I'll definitely count that
as a success, because that's been a sticky point for large numbers of people.

NF: Within the Debian project, or without?

BR: Largely within, but in part without. The veil of mystification keeps people from getting involved
in ways that they could help, simply because they can't see a fit for their skills.

NF: Will one year be enough to accomplish that goal?

BR: Probably not. But I would like to be the most visible DPL in that role -- not in terms
of going to the most trade shows, but in doing the most in helping my fellow Debian
Developers understand what it is to be Project Leader, the types of decisions that get made,
and, really, to help the Debian project understand what it needs in a project leader.

NF: Is there anything else that you would like to say?

BR: Just that it's very exciting and humbling to have been elected. It's definitely a change of perspective to achieve
something you've worked for for many years, especially when you've just been so pessimistic about it.

NF: It seems like an ideal integration of your daily work at Progeny and your free software activities.

BR: It is. I think that my non-Debian hobbies will suffer even more than they have. But I don't think that I'll have too many
regrets on that front.

Bruce Byfield is a freelance course designer and instructor, a technical journalist, and a regular
NewsForge and ITMJ contributor.

Click Here!