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

Re: Debian needs more buildds. It has offers. They aren't being accepted.

Scripsit Will Newton <will@misconception.org.uk>
> On Thursday 19 Feb 2004 10:30 pm, Henning Makholm wrote:

> > the DAM? There has to be someone in that position, someone with the
> > guts to sometimes say no where the AM has said yes. And it is
> > desirable that, if possible, it is the same person each time, for
> > consistency in the processs.

> I disagree. I don't think it is desirable that it be one person and
> they be "consistent". I belive that will risks a monoculture of
> developers. If, as seems to be have been suggested, the DAM has a
> problem with impatient people, then we get no impatient people in
> Debian, regardless of ability.

Well, it certainly doesn't *seem* like Debian has any shortage of
impatient people...

Remember to optimise for the common case. The common case is that
someone who gets thumbs-up from his AM is indeed qualified to become a
member. That is the case that needs to be efficient, and it's not
going to be efficient to double or triple the number of people who
need to read through the application's proceedings and form an
opinion. As one who has recently went through the process of answering
the standard question sets, I can attest that it takes forever to
*write down* the answers, even if one knows most of them in
advance. Reading them is of course somewhat faster, but I would be
surprised if it is possible to form a qualified opinion about an
application in less than a couple of hours. And writing a
well-reasoned account of one's deliberations for public consumption
could easily take an hour more than making the decision in the first

The less common case, where the material collected by the AM does
*not* convince the DAM of the applicant's suitability, is where it
becomes relevant to include committees. And that is indeed what
happens, at least if the information on
is to be believed. If it is not to be belived (i.e, if the DAM goes on
a spree and starts using "ultimate" rejections regularly), I am sure
some of the committee members are going to notice and cry havoc.  It's
a rather big committee, and it would be extremely difficult to say
with a straight face that they are all in the same cabal.

The final possibility for abuse that remains is if the DAM simply
stalls an applicant without rejecting him outright. However, unless we
see concrete evidence to the contrary, I prefer to believe that
stalling at the DAM stage is caused by hat overload rather than malice.

> 3. A public mailing list where all deliberations are archived (or even a 
> debian-private alike list if that is too much).

I think that applicants who are going to be deservedly rejected would
much prefer not to have detailed explanations of why they suck
available to everyone who cares to type their name into Google.

> 4. If the a committee member (or DAM) has a personal problem with a
> NM, they should declare a conflict of interest

I'm fairly sure that most people in Debian who have personal problems
with someone else genuinely belive the problem to be an objective one
with the behaviour of that someone else.

> If it takes an email and someone else to add the key, so what?

It means that that email is going to have to propagate to the top yet
another person's to-do-list before the applicant has a key in the
keyring. And the perceived problem that most people are complaining
about is *not*, as far as I can tell, that people are being rejected
unfairly -- it is that the applicant waits too much.

> I think the monthly flamewar should be a hint.

It is a hint that there is at least a communication problem somewhere.
I am not convinced that it hints of a problem that can be mitigated
by adding more red tape to the process.

Henning Makholm            "My fate? Servitude to the Embodiment of Whoops."

Reply to: