Petteri Räty wrote:
Mike Auty kirjoitti:
Petteri Räty wrote:
Defining required amount of activity for ebuild devs. I would like us
to raise the required amount of activity for ebuild devs.
Given that the low number of developers is ranked as our number one
problem in Donnie's informal survey[1], taking any kind of action
against infrequently-committing developers is likely to reduce the
number of devs we have, and potentially make the problem worse.
I agree with the above point.
Also, as I recall, both Pettery (betelgeuse) and Denis (calchan) have
stated before that we no longer have any queue of people waiting on
recruiters to join Gentoo. I'm not seeing an avalanche of new blood
entering the distro, so I'm wondering where we want to go.
If we keep going the route of the last months, I wonder how long it will
take until we get under 150 devs. I don't think this will benefit
anyone. Furthermore, the trend in the last months was in large part the
result of finally retiring people that had been slacking for a long
time. This proposal could (would?) lead to sending away people that
still do work, albeit at a slower pace or on bursts.
What benefits are you aiming to get from the suggestion? I can think
og keeping the books tidy and reducing management time required to
maintain the devs. Are there others I've missed? If they're worth
the cost/effort involved with putting someone through the dev tests
and getting them trained, then it seems a good idea, but otherwise
probably not...
Mike 5:)
[1] http://dberkholz.wordpress.com/2008/02/21/redux-gentoos-top-3-issues/
If you can't manage weekly commits, you can't respond to security issues
either. This means that you should have devaway on.
As others have commented, I don't agree with this point. Also, you're
forgetting we have quite a few people working on this project and that
we have many different roles.
Although you're talking about ebuild devs only - so doc devs, infra and
forums staff are exempt from this rule - you're assuming (asking?) that
all people with access to gentoo-x86 are package maintainers and do a
few, regular commits to the tree. As others have said, that assumes
people keep more than a few ebuilds and that those packages require
constant attention.
Recalling previous discussions about work on gentoo and some of the
existing roles, what will you do to AT folks, release members or QA
members? Are they also obliged to do a weekly commit to keep their
"privileges"?
Finally, I thought the whole point of removing access to infra boxes
(which is the end result of retiring a dev), was a concern with security
and not a way to get rid of people - with the exception of
administrative action by devrel.
We've been having a few discussions about the future of Gentoo for some
time and people have shown different goals and views on its future and
on how to get there. One of the views seems to be that we need (only
need?) an "elite" of super-devs that do daily (hourly?) commits. I have
nothing against people that can give so much to this project, but I
don't think it's reasonable, desirable or healthy to expect everyone to
be able to that level of commitment. Also, wasn't this distro at one
point all about community? I don't think raising the commitement level
helps to involve people and as William (wltjr) pointed out shouldn't we
be more concerned with quality than with quantity?
I understand and agree that ebuild devs should keep a minimum level of
work to justify their access to the gentoo-x86 tree. I would also like
to have a few devs that can do major commits (although commit sprees can
have their own problems), but I think there's still a place in this
distro for people that want to maintain a few packages, that want to do
AT work, that care with the QA of the tree or that work on releases.
These people shouldn't be sent away, just because they can't keep with
weekly commits (not enough work or time?) or because they work in bursts.
As a final thought, I think this point is a tangent to the old debate
about tree-wide commit privileges and or the scm of the tree. Afterall,
if gentoo-x86 was a git tree and or we had acls in the tree, I don't
think we would be having or would need to have this argument.
Regards,
Petteri
--
Regards,
Jorge Vicetto (jmbsvicetto) - jmbsvicetto at gentoo dot org
Gentoo- forums / Userrel / SPARC / KDE
--
gentoo-dev@lists.gentoo.org mailing list
