Introduction
I'm a 28 year old man. I'm married and have no kids yet. I'm self-employed and
I decide alone how I spend my time.
Concerning Debian, I've always been interested in organizational issues and in
quality assurance. To give more precise projects: I've helped setting up Alioth and I co-administrate it currently.
I created the Package Tracking
System and still maintain it with the help of a few others. A long time
ago, I also pushed forward the sponsorship
principle. I did my share of packaging, bug squashing, NMU,
perl/python-policy creation, debian-cd maintenance and stuff like that all over
the 8-9 years where I have been involved (1998-2007).
I already ran once for DPL in 2002. Please check the platform that I wrote
at that time. Check out how many of my ideas/projects were right and have
been implemented since, and how I consistently implemented several of the
projects that I announced.
Why do I run for DPL this year? Only because...
I want a DPL board
I think that I have a good understanding and a good experience of the project
as a whole. I would like to use that experience to serve the project, but would
I have to handle alone the duties of such a demanding role, I would not run.
I'm sure that I'm not alone in that position. I think the time is over to put
too much hope in a single candidate each year.
I'm really convinced that working within a DPL board is the best way for me to
serve the project and make constructive propositions. I'm in no way perfect,
and I make mistakes, but I'm used to working with others (even when they don't
agree with me), and the DPL board needs to exist because of this: I always need
the advice of other trusted members to enhance any proposal so that it can be
broadly accepted. I don't need to fear antagonisms within the board, because
I'd rather drop an idea when I see that 3 members don't like it instead of
proposing it broadly and generating yet another useless flamewar.
Believe me or not, I'm convinced that we can
agree on some basic
level and build from there.
How could it work? I outlined it in a recent mail
to debian-project and I asked for feedback. I plan to delegate all the
powers of the DPL to the board with a set of rules defining how the board can
take decisions. After that I shouldn't have to directly use my DPL powers
(except if the board needs fixing), and I expect to play an active role in the
board making proposals when needed.
All board members would be added to the alias [email protected]. Near that a
publicly archived discussion list restricted to the board members will be
setup. That list would be the main tool for the board. Whenever a board
member feels the need for an official decision, he makes a proposal which
is then discussed on the list.
Some time after, the proposer decides
that enough discussion has happened and he asks for a vote. This vote
lasts for one week and the quorum is very small. The vote is usually a simple
yes/no/further discussion. But the proposer may request a Condorcet vote
if that's what is most appropriate according to him.
The board should represent the project in its diversity at a smaller
scale. This should enable discussions by proxy
where the discussion stay
civilized because the board has been chosen with that objective in mind.
The membership of the board should be chosen with a goal of maximizing
the ability of each developer to relate to someone in the board so that he
can express his point of view with confidence that someone will
understand him and represent his opinion.
What would the board do? Whatever the board members think is needed. They have
the DPL power and I expect them to use it wisely to pursue any beneficial
Debian activity.
If possible I would like to work with this board:
- Bdale Garbee <[email protected]>
- Steve McIntyre <[email protected]>
- Jeroen van Wolffelaar <[email protected]>
- Enrico Zini <[email protected]>
- Josip Rodin <[email protected]>
- Steve Langasek <[email protected]>
- Wouter Verhelst <[email protected]>
- Sam Hocevar <[email protected]>
The first four are unconditionnaly ready to serve within a DPL board. Josip
would rather support an elected board like he
described on debian-project. The others prefer to reserve their decision
until the vote is over. Steve Langasek obviously wants to wait until Etch is
out. Wouter and Sam would probably prefer be involved if the project decides
that having a DPL board is a good idea. And they also don't want to indirectly
endorse my candidacy.
This board will probably be extended by one german-speaking developer and another one from asia/australia. I couldn't get the required agreements in time for this platform, so it's left as first exercise for after the end of the election.
My personal projects / opinions
Of course, if I am elected, the ultimate decision on my projects will be up to
the board. Still, it's worth telling a bit more in which direction I'd like to
push the project.
Transparency and communication
I've read my share of flames over the years and it's getting boring to see that
we have some recurring problems that could be mostly avoided with a couple more
mails coming from responsibles behind various role addresses. Obviously they
have their reason to stay silent most of the time and I believe this is
directly related to the (bad) communication habits that we developed within the
project.
My experience with Alioth proved to me that giving news in public will bring
you some unneeded criticisms. Nevertheless, I continue to give news and handle
much stuff publicly because I believe it's required for the long-term health of
the team. Without this we couldn't have recruited two new volunteers.
Alioth is now used to maintain many packages and develop important part of our
infrastructure, so the service is important and as closely watched as other
DSA-maintained services. But it looks like we avoided most of the criticism
while handling it.
I want to build on that experience and write a set of guidelines for internal
teams
that would be widely approved by the project, and then, over time, work
with the various teams to help them meet those objectives.
Use what we package and package what we use
This is not really a project, but a policy that I'd like to promote. There are
too many cases where we're using stuff which is not packaged in our main
repository. In general we have good reasons for that: because it is
Debian-specific and/or a hack. Examples that come to my mind: the
DSA-maintained ssh, the PTS (yes, I did not package it), the difference between
sbuild on buildds and the sbuild package, etc.
I think we should aim at always using our Debian main
packages because that
would require us to think a bit more globally and adapt our tools so that they
can serve others too. We have many derivatives but they don't always use the
same infrastructure as we do, partly because of that. Another advantage of
having packages is that it naturally offers the BTS interface that can be used
to report problems, ideas for improvements... and this in turn is useful for
volunteers who want to help out some specific team and who are looking for
something interesting to do.
Empower people to do (useful) stuff
Historically, we have numerous examples of people who have done stuff alone
without any approval from the project/team in charge: Anthony wasn't ftpmaster
when he wrote britney, I pushed for Alioth because I thought it was a good idea
and DSA never approved nor blocked anything, the same goes for the testing
security team.
If some people feel they are blocked in their work by a current problem and are
willing to help, I'll do whatever I can to provide them the required resources.
Alternatively I will come up with innovative ideas to go around the problems.
Let's take the example of the lack of DD-accessible alpha porter machine. We
could imagine a service that would let third parties offer access to their own
machine. The Debian Developer who needs access to such a machine would request
it on some webform. The machine owner would just have to install a specific
package which would export three chroot and auto-update the access list
(including public SSH keys) every 15 minutes from a central trusted place. The
machine would also export some items of description (contact info, RAM,
available disk space, CPU type, ...) that can be used by the developer to
select on which machine he's likely to be able to do his work.
Support the occasional contributors
We strive to be the universal operating system and I fully share this goal. But
for this, we need to leverage the power of the many occasional contributors.
Some recent discussions have shown that we can do better at this: we could
provide some better overview of packaging related changes to the hundreds of
packagers that spend only a few hours a month on their (usually small)
packages. Announcing policy changes on debian-devel-announce has been proposed
and is likely to be done. I also think that the idea proposed last year by
Anthony Towns should be implemented: after some successful rounds of
sponsorship, we should be able to give upload access for that specific package
to its maintainer and we shouldn't require him to go through NM if he doesn't
want to because he's not more broadly interested in the project. Of course, we
have to put some checks to avoid any loss in quality.
Why vote for me?
Why not? It looks like lots of my projects have had a positive outcome. Let's
try some more!
Rebuttals
On my own platform
I have voluntarily focused on a limited list of projects but there's much more to do, in particular concerning Debian's outside communication. I'd like to use the board to draft some documents explaining the relationship of Debian with the various players from the free software world (see the original idea from my 2002 platform). Also we have some good communicants within the board and I'm sure that this will lead to more visibility for Debian.
Wouter Verhelst
I agree with most of what Wouter said. His plan to discuss first and
decide next, is the only sane thing thing to do. But this information
gathering
step is best done within a team because it provides a better
coverage of the various inter-personal relationships that are into
play.
If I'm elected, I hope that he will confirm his participation in the DPL board and that he will pursue all the goals that he outlined in his platform.
Sam Hocevar
All of Sam's goal are interesting. I share most of his analysis of the situation. However, as an independant DPL I fear that he would be too much isolated to effectively interact with all the internal teams.
If I'm elected, I hope that he will confirm his participation in the DPL board and that he will pursue all the goals that he outlined in his platform.
Steve McIntyre
Steve's platform doesn't give many concrete projects but rather a global direction that he wants to follow. I welcome any initiative to solve the problems that he outlined.
I'm glad that he accepted to be part of the DPL board.
Anthony Towns
Anthony recognizes the need for more assisting leaders
to share the
load
. He even explains how being involved in DPL activities helps you
being more effective as full DPL later.
Clearly being DPL gives you access to more information than the average DD and this information is crucial to make Debian evolve without (too many) disruptions. I think this information needs to be shared by as many wise persons as possible and this is the intent of a DPL board.
I like his choice of topic-keywords but most projects outlined do not
require to be DPL. I'll gladly provide him the required support so that he
can complete his
project of supporting non-DD maintainers. Having ambassadors
is a
nice idea, but with a board, we already have many co-DPL which are
de-facto ambassadors of the project. I'm still open to the idea
however.
Aigars Mahinovs
His big plans won't come into reality simply because he has been elected. The only Debian-specific project is so ambitious, that it's unlikely to ever be completed. However I'd definitely welcome new tools to make collaboration with derivatives easier, but this won't happen if it needs a big-bang change like he suggested.
Gustavo Franco
There are so many ideas in his platform, that he should have structured it and prioritize them. I lack the reasoning behind this large selection of projects/ideas.
Simon Richter
Simon sees the DPL as someone who must participate in all important discussions with the goal to promote an intermediary, reasonable position. This is a difficult task because you inevitably favor your own opinion. That's why I prefer having the discussions at a smaller-scale within a board representing Debian's diversity. The goal is the same, but it's done differently.
His platform only tackles this single internal communication problem. I lack a broader view of the direction where he wants to take Debian.
Sven Luther
His platform was not online when I wrote the rebuttal.