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

Re: Using money to fund real Debian work

On Wed, Oct 11, 2006 at 03:00:02PM +0100, Ian Jackson wrote:
> The RMs personally are in a privileged position for requesting
> funding: 

The RMs didn't request funding.

> their role within Debian is critical for the whole project;

Plenty of people run unstable, and it's often been mooted that we
shouldn't bother with stable releases at all. Debian would continue if
we didn't do stable releases, just as it would continue if we didn't
support some particular architecture, or some particular packages: it
would be less amazing for the loss, but that's all.

> it is not easy for someone else to come and learn how to do it; 

Steve started learning how to do release stuff in March 2003, following
the request for help I sent out [0] and was given the title of "Release
Assistant" in Aug 2003. In Aug 2004 Steve and Colin took over as joint
Release Managers (that's not much more time than it often takes to go
through new-maintainer, note), and in September 2005 Colin resigned as
RM, with Andi becoming an RM in October 2005 [2]; in July 2005 there
had already been a call for new assistants [3], which led to Marc
Brockschmidt, Luk Claes and Adeodato Simo being appointed RAs earlier
this year, and getting the extra special release permissions.

  [0] http://lwn.net/Articles/25154/
  [1] http://lwn.net/Articles/45232/
  [2] http://lists.debian.org/debian-devel-announce/2005/10/msg00004.html
  [3] http://lists.debian.org/debian-devel-announce/2005/07/msg00009.html

> it is
> not easy for someone else to get permission to do the work instead;

IME, all you have to do is volunteer.

> If I decide I can do the RM job better than the RMs, and my pet
> benefactor agrees with me, I still can't get paid to do it unless I
> can get the existing RMs and/or the DPL to agree.  

That's not true -- we've already had an instance of that happening with
Mark Shuttleworth as the benefactor and Jeff Waugh amongst others as
the guys who think they can do the RM job better than the RMs.

We've had other, unfunded, examples of that within Debian too, in the form
of the "volatile" and "backports" archives that implement a different
stable update policy than the Stable Release Manager used. For that
matter, we've also had the debian-amd64 release of sarge for amd64,
which was done not only separately to regular release management, but
separately to the main archive too.

> And of course there
> is a (financial, now!) incentive for the existing RMs not to agree. [1]

There's also a (financial, now!) incentive for them to look at the
ideas the new RM volunteers have had and implement them themselves if
they work better, so as to encourage more people to see that release
management is working well, and be willing to contribute in the future
to its continued success.

Furthermore, what you're arguing for is that benefactors should not be
involved with people who volunteer to do the RM job in the first place,
so even at worst, you're still better off with the RMs having a incentive
(financial or not) to ignore your contributions, than having the project
as a whole develop a stated policy against accepting contributions
enabled by a benefactor. Better a chance to contribute than a certainty
that you'll be ignored.


Attachment: signature.asc
Description: Digital signature

Reply to: