Request: please consider clarifying the project's position on Spins

[Christoph Wickert]

Am Mittwoch, den 01.12.2010, 13:42 -0800 schrieb Jesse Keating:

> There is also the question of who /tests/ the things that are produced.
> 
> Even though I've stepped down as Fedora releng, I think it's pretty safe
> to say that anything that is produced and published "by Fedora" needs to
> go through the rel-eng department.  

We did discuss this with Jared during the EMEA FAD and we came to the
same conclusions:
     1. Spins need to be properly tested.
     2. Spins need to be properly produced by rel-eng.
     3. These two are often the bottle neck.

> So having those that want to create spins join releng makes sense.

Correct me if I'm wrong, but rel-eng is the only group in Fedora that
has no instructions how to join in the wiki. For most of the Fedora
contributors they act as a block box that is constantly working and
magically every 6 months it a release falls out of it.

IMHO large parts of this is a communication problem. rel-eng fails to
communicate what they do and how they do it. Just a few examples:
     1. As a spin owner I was not notified in when F12 test composes
        where available for testing. Spins SIG was not notified ether.
     2. There was a problem in the F13 LXDE spin that required a respin
        (but no updated packages) of the image. It was not done as
        rel-eng did not give feedback in the corresponding bug report.
     3. I have been fighting for the multi desktop DVD and suddenly
        someone brought up to my attention that there is an image at
        http://alt.fedoraproject.org/pub/alt/mega-dvd-test/
        Who produced that image? Why didn't he bother to tell me?

I don't want to point fingers at rel-eng because I think they simply
lack manpower. What I want to point out: When it comes to communication
I think we are no longer dealing with "low level" problems as Toshio
called them. It's on the board as the primary Fedora body to define a
workflow how other groups within the project should interact and
communicate.

Another problem we are facing is the policy problem that Adam already
mentioned. Are spins valid blockers for a release? Do they get the same
amount of testing? The lack of a proper policy also needs to be
addressed by the board I think.

I'm pretty sure that the spin SIG or the individual spin owners are
happy to provide input on all these questions, but in the end policy
decisions need to be made by the board.

Regards,
Christoph