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

Re: Developer Status

Hi Luk,

2008/10/23 Luk Claes <luk@debian.org>:
> Raphael Geissert wrote:
>> Right, but do the members of the NMC cover the wide variety of
>> programming languages?
>> or what kind of review are they going to do? just packaging stuff? if
>> it is just the latter it would be much easier and faster to send a RFC
>> to -mentors and let people scream out loud.
>> And please note that I said "QA side", with which I didn't mean to
>> refer to the QA group, but to a variety of people who know what to
>> look at and how to do it; not a random AM who happens to have already
>> completed doing a review process successfully (which actually doesn't
>> guarantee that the AM is competent enough, as the usual NM process
>> consists on sending the templates and later reviewing the responses).

First of all I would like to apologise as it was not my intention to
* offend/criticise anyone's work,
* mean that NM is all about questions, as it of course isn't.

I believe that impression was caused by my oversimplification of the NM process.

> You're very wrong here. The AM's job is to review if someone would be
> capable of being a good Debian Developer. Reviewing responses to the
> templates is *not* the main job. Have the prospective DD learn things;
> get the prospective DD think and search before answering; and reviewing
> actual tasks and skills by reviewing the prospective DD's packages next
> to possible other 'tasks' takes most of the time.
> It's not at all about a questionaire where you only have to tick the
> right answers because that would defeat the spirit of the process. For
> many applicants it takes a long time because they think it's just a
> questionaire...

Getting a little bit back on my point: what I wanted to mean with my
explanation of "QA side" is that:

* nobody is perfect,
* not always packages are checked as rigorous as they, at times, are at -mentors

So, my proposal is more or less like this:

1.- Keep everything in public MLs
2.- Assign a member of the NMC to the "case"[1] who will of course be
willing to work on it
3.- Publicly request[2] people to review the package[3], although the
NMC member will also do it on her own.
4.- Every review should be handled[4] by the NMC member and a decision
be made (also refer to [1]).

And as an extra note: any NMC member should be able to take over the
request at any time in cases such as when the review/process is not
objective (i.e. there are personal feelings).

[1] This member will take care of the case/request and will write a
brief, but explanatory, report when making a decision.
[2] Teams, co-maintainers, etc should probably be CCed.
[3] No review should be disregarded, no matter whether it comes from a
D<something> or from a D<nothing> as any input should be considered
[4] Any extra investigation needed originated by a contributed review
should be made before any decision is made.

Does that sound better? comments? suggestions? positive criticism?

Although I'm not still very convinced I at least want to propose
something that, if approved, would make me happier than the
originally-proposed process.

> Cheers
> Luk

Raphael Geissert - Debian Maintainer
www.debian.org - get.debian.net

Quentin Crisp  - "If at first you don't succeed, failure may be your style."

Reply to: