Proposed release criteria revisions

[James Laska]

On Fri, 2011-03-18 at 15:14 +0000, "Jóhann B. Guðmundsson" wrote:
> On 03/18/2011 02:27 PM, James Laska wrote:
> >> I disagree with this. A lot of what we want tested needs a graphical desktop.
> >> >  If we were to ship an Alpha without a working desktop, we wouldn't get much
> >> >  feedback (other than that the desktop doesn't work).
> 
> 
> That's mostly cause we are working on outdated thoughts and dreams of 
> the project which the project and the community has outgrown long time 
> ago by shipping an "Default" in any shape or form .( From my perspective )

So we should define criteria that apply to a project model/workflow that
doesn't exist?  I'm all for having criteria that work for our project,
which I believe this thread is addressing.

> So I can get a bit on the page what is the feedback you general expect 
> surrounding alpha and from whom?
> ( Rawhide users/reporters should know it's broken in the first place and 
> provide feedback )
> 
> Shipping "Alpha/Beta" is only a marketing gimmick anyway and it outdated 
> the day we finally ship that stuff..

You forgot to add ... "in my opinion"  ;)

Marketing is certainly one important angle in *any* release milestone.
I wouldn't say it's only driver.  If we want to debate the merits of
release milestones, that perhaps should be a different thread.

> > I agree with Bruno.  The suggestion conflicts with one of the stated
> > objectives of the Alpha, "Test accepted features of Fedora 15" as this
> > would prevent testing of any desktop-related features.
> 
> If I'm not mistaken the the feature process entirely optional so tying 
> the "testing" of features into our release cycle so closely does not 
> make any sense to be perhaps someone is willing to refresh my memory by 
> explaining it to me why we did that in the first place?

Read the lists/wiki for the history.  The feature process has it's
flaws.  What adjustments do you recommend?

> > That said, if we ever have SIG-specific criteria, the proposed change
> > would be perfectly suitable criteria for the Server-SIG.
> 
> It is time for us ( QA  ) to get more in times with the evolution of the 
> project and start considering SIG specific criteria or better yet a 
> solution more consistent with and built upon other things we have been 
> working like critical-path.

As I said in the previous mail, I don't disagree with the idea of
SIG-specific criteria.  The beauty of this behind-the-times team is that
you can propose an idea, and if it's good+scalable+effective, it will
stick.  My caution with SIG-specific criteria is that I believe there
are policy issues beyond just criteria definition.  We are hijacking
Adam's thread, so perhaps this would be best started as a new topic...
describe your vision of SIG-specific criteria, what is success in that
world?

Thanks,
James
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
Url : http://lists.fedoraproject.org/pipermail/test/attachments/20110318/f784aa48/attachment.bin