Role of the Sponsorship Queue
persia at ubuntu.com
Thu Mar 4 11:06:26 GMT 2010
Daniel Holbach wrote:
> I too remember when we didn't have a Sponsoring Queue and it's not where
> I want to go back to. :-)
> If we didn't have 2000 bugs with patches, I personally feel we could
> very well live without a sponsoring queue and just have a look at the
> list of incoming patches. As I see it right now the sponsorship process
> helps us to not fish in the dark and it ensures that we're looking at
> bugs where the problem is well-understood and deal with genuine patches
> as opposed to random files (we could discuss further measures to make
> sure the quality is good enough). Also do 'workflow bugs' go into
> Sponsoring if necessary.
> The process might not be perfect but, given the reality we're living in,
> addresses the biggest problems we have.
I feel very strongly that if the biggest problem we have is that
there are 2000 patches laying around, then we need to fix that by
going and dealing with the patches. Complicating the process by which
a patch is reviewed is not the way to make more patches get reviewed.
Developers asking non-developers to go review patches before
developers do is likely to just reduce the average quality of response
to patch submitters : If developers are good at patch review, we're
removing our best team from the set of reviewers. If developers
aren't especially better at review, we're telling everyone else they
are second class which can't help motivation. If 2000 is the correct
number, then that's only about 15 patches per developer. If not so
many developers are active, that's maybe 30. If everyone reviews one
patch a day, we'll be done somewhere between a two and six weeks,
depending on how many of us there are.
If we're not serious about fixing the problem of unreviewed
patches by just doing so, then we're not helping either ourselves or
anyone else by creating an extra process step that creates the
impression that it's hard to submit patches to ubuntu and filling a
sponsorship queue with all sorts of stuff so that sponsors can't be
confident of motivated actice submitters. I'm not sure how this is an
improvement of the experience for either group of people, and I'm sure
it's demotivating for anyone who works on packages for which they
don't (yet) have upload rights.
More information about the ubuntu-devel