[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Question for Andreas and Branden

On Fri, Mar 04, 2005 at 08:16:34PM +0100, Guido Trotter wrote:
> First of all I want to thank you for running as DPL.

It's never been an easy decision to make for me -- and it gets harder every
year -- but your support makes it easier.  I'm sure Andreas similarly
welcomes your support.

> In order for us to produce a useful ranking, though, I think it's it
> would be nice from you to elaborate a bit on what exactly things would be
> different if one or the other gets elected. Of course only the winner
> will hold the official position of "the DPL", be formally responsible for
> all the decisions, be crowned at the debconf (sic, I'm so sorry I won't
> probably be able to be there this year), etc.

It's difficult to make predictions of the nature.  Andreas and I have
enjoyed a good working relationship for years.  I think the differences
will be seen largely in terms of which areas are specifically tackled
personally by the DPL, and which are delegated to another team member.  In
principle this should be much of a difference, but if you're looking for
a way to decide which of us to rank higher, then I recommend that you rank
more highly the candidate that identifies more areas of concern that you
share, and whom you think will be more effective at grappling with those
areas.  The DPL directly answers to the developers in a unique fashion, and
will experience a unique burden of office.  The team approach is engineered
to mitigate the negative aspects of that, but I don't think it can take
away from the sense of enthusiasm a person enjoys in being elected by their

As far as a crowning ceremony goes, I'm not sure how enthusiastic I'd be
about such a thing.  I'm sure it was undertaken for fun last year (probably
on a lark, as I don't recall any such ceremony from DebConf 3), but unless
it is irreverent in the extreme, I don't think it would suit my personal
style very well.  Given my history with the project, a "roast" might be
more appropriate.  :)

> What I would really like to know is if you think that the decisions and
> the actions of the team you're proposing will be the same if either of
> you gets chosen, or that things would change, even slightly, and in this
> case which would. I think knowing this is really important in order to
> avoid people "flattening" both of you in the same position in their
> ranking, or just rank you based on personal sympathy, rather than
> technical/political position.

Andreas and I do have our disagreements about things, though they are
friendly.  I think this is a plus, not a minus -- there is no point in
running two candidates if they *really are* clones of each other.

Here's one concrete point of disagreement.  Andreas is pretty earnest about
"changing the culture" of Debian, and has brainstormed some ideas along
these lines, mostly having to do with punitive enforcement of courtesy
standards on the mailing list (much like Anthony Towns proposed[1]).

I've cautioned Andreas against such proposals, because they're not only
inherently subjective, but they put really large burdens on the people
appointed to enforce the policies.  Not only do they have to have code and
procedures in place against attacks; (Wanna shut someone up?  Forge mail so
as to fool the list censors into banning him or her! Recall that not even a
majority of posts to debian-devel are GPG-signed[2].); they'll have to have
tremendous amounts of time just to sift through all the traffic on the
high-volume lists; and they'll easily become the targets of complaints
about unfairness, much as the archive administrators are now[2].

To be fair and just, there would need to exist an arbitration body to hear
appeals from people who felt they were dealt with excessively harshly.  We
could do without one, of course, in which case these things will just roll
downhill into the DPL's lap, having nowhere else to go.  Given that the
control-bot ban list[4] in the BTS was implemented with apparently no
review process whatsoever, in that bans were instituted without
consultation with or notification of other BTS admins[5], and there is
*still* no documented procedure for how such bans are instantiated,
reviewed, appealed, or lifted, I'm not sanguine about our odds for doing
better at instituting a committee to ban people or discussion threads from
the mailing lists.

My take on this would be, assuming we could find people willing to serve as
list sifters, I'd try to encourage them to be "beat reporters" for DWN.
Joey Schulze needs all the help he can get[6], and some of the critiques of
his work in that area that have been seen here on -vote recently might be
mitigable.  If people find Joey's coverage of -legal to be inadvertently
sensationalistic, for example, then maybe what's needed is someone to help
relieve the burden.

With better digesting of list traffic, we can better separate the wheat
from the chaff.  Any such selection process involves trust, and risk that
someone's subjective standards won't match your own, but the advantage to
my proposal is that it's subject to almost arbitrarily wide peer review.

This doesn't mean Andreas has bad goals or bad ideas -- it just means I
have a friendly disagreement with him.  If he is elected DPL, I will advise
him on this issue to the best of my ability, as I will with others, and if
he doesn't persuade me, I will exhort him to try my approach first.  If I
am elected DPL, I will ask him and the rest of the DPL team for commentary
on my proposal.  I reserve the right to put a question to the entire
community of developers, and expect to do so for the contentious issues.
The monthly updates I pledged to make as DPL will be, I suspect, a good
forum for putting forward such interrogatories.

I think each of us will tend to emphasize the problem points we identified
in our platforms (my problem points actually come from my 2004
platform[7]).  Which set of goals would you rather see worked on?  Rank the
better fit higher.

> Thanks for your attention and your time!

Thanks for your question.  Please let me know if I have not addressed it to
your satisfaction.

[1] Message-ID: <422852C6.1090306@azure.humbug.org.au>
    Message-ID: <42274908.1050803@azure.humbug.org.au>

[2] According to Mutt's "~g" pattern and my mail folders, of 3663 recent
    messages to debian-devel, 972 bore a GPG signature.

[3] Of course, being an archive administrator himself, perhaps Anthony
    feels it's high time someone *else* played project whipping boy, and
    this risk is actually a feature, not a bug.  :)

[4] In speaking with one of the BTS admins, I've learned that there has
    been one technological change and one process improvement.  The Debian
    BTS apparently now uses procmail rules instead of a passionately-named
    Perl variable to keep track of bans, and bans now have comments next to
    them.  The example I was shown looked like this:
    # Banned by X 'til 2004-01-26 for repeated reopens on Bug#Y

    [I obscured the name of the banner and the bug number because I don't
    think they're relevant to discussions of general method.]

[5] Message-id: <20031105071251.GA15771@deadbeast.net>

[6] http://www.nl.debian.org/News/weekly/contributing
[7] http://people.debian.org/~branden/dpl/campaign/2004/platform.xhtml

G. Branden Robinson                |     If God had intended for man to go
Debian GNU/Linux                   |     about naked, we would have been
branden@debian.org                 |     born that way.
http://people.debian.org/~branden/ |

Attachment: signature.asc
Description: Digital signature

Reply to: