Freerunner's Future / How will the kernel be maintained?
Carsten Haitzler (The Rasterman)
raster at rasterman.com
Wed Jun 3 09:37:21 CEST 2009
On Tue, 2 Jun 2009 19:52:59 -0300 Nelson Castillo <arhuaco at freaks-unidos.net>
said:
after having run e for years... i'll drop my advice here. it's simple. trust
people. let people send patches - i they send more than a few over a short
period, offer them commit access. people who ask for commit access get it. if
people commit and screw the kernel they of course get warnings and public
beat-downs. everyone knows who commits what and when - have a git commits list
that sends emails on every commit with the diff's. people who cause too much
trouble simply have their commit access revoked and are of course on a
blacklist. this happens incredibly rarely, so it' better to ask for forgiveness
than seek permission. of course this requires being willing to give up
direct control. you play for of a parent-figure just watching the kids play and
do heir thing - guiding every now and again and spanking when they really do go
astray. it reduces overhead significantly.
> Hello there.
>
> As announced by Sean Moss-Pultz in the Community List the GTA02 future
> will depend on the community.
>
> I think this is already the case on the kernel side thus things will
> not change much here (things already changed in March). All the
> involved kernel experts are community members now anyway. In a
> previous conversation with Werner we talked about opening the process
> even more by not having a (single) maintainer and instead allowing
> more people to commit.
>
> Now the question is how to do it.
>
> We know this will be volunteer work thus we need to find people who:
>
> - Are willing to do it
> - Are willing to check/test patches before committing them
>
> It would be better if this access was given to people who have
> submitted correct patches in the past or who send nice patches in the
> future. I will be around with lots of time for about three weeks thus
> I can help with tasks. I think we can make a smooth transition form
> the centralized OM model to a more open one with no problems.
>
> Cheers,
> Nelson.-
>
> PS: See: http://lists.openmoko.org/pipermail/community/2009-June/048980.html
>
--
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler) raster at rasterman.com
More information about the openmoko-kernel
mailing list