Re: Question for all candidates
Steve McIntyre wrote:
Do we actually need a DPL? Would we be noticeably worse off without a DPL?
ObVious: We'd be violating the constitution not to have one; if we
refuse to elect one, we'll just have Manoj and Ian act as the DPL until
we work out what we want.
The general answer is, in my view, that we need some way of making
decisions that can't be made by a single package maintainer, such as
"who should be the maintainer of this package?" Even if we do that by
writing some scripts, and forbidding anyone to change it -- we have to
come up with /some/ answer; and the question is then what is the *best*
answer?
Options other than a single DPL responsible for the entire project is to
have a committee much like Jeroen's Scud team. But committee's are
really hard to get right: sometimes they're too diverse and their
recommendations are a hodge-podge of incompatible ideas that are overly
complicated and unworkable; sometimes they end up having only one person
do any of the work, and then handicap that person with the formality of
approval by the rest of the committe; sometimes they waste everyone's
time in unproductive meetings; sometimes they get in the habit of
debating things internally and poeple who aren't on the committee don't
find out about what happens for months, and may never know why it happens.
Those are just fundamental properties of committees and teams in
general; while they can be absolutely brilliant when they're working
well, when they're not they can be truly, truly awful. I think it's
definitely an approach to work on -- but given our experience with the
technical committee's responsiveness or the SPI board's effectiveness, I
think we should be fairly wary of putting committees in leadership roles
within Debian.
Other options are to just have delegates doing the work and answerable
mostly to themselves; which some would argue is already the case. But
most of the people who do then go on to argue about how horrible this
is, and how it's destroying Debian; some sort of oversight really does
seem useful.
Direct democracy -- ie, solving problems by GR -- is another solution; I
think the level of debate and formalism that involves tends to be overly
political in all the bad ways, and suspect it's not something most of
the participants in Debian have the patience for.
Heck, I'm pretty impatient with the nine weeks that get spent on DPL
elections each year, personally...
Cheers,
aj
Reply to: