Critical Path Wranglers - Draft Proposal

Adam Williamson awilliam at
Tue Mar 9 03:06:04 UTC 2010

On Mon, 2010-03-08 at 17:38 -0800, Jesse Keating wrote:
> On Fri, 2010-02-26 at 23:55 -0600, Adam Miller wrote:
> > In the last QA meeting it was discussed that we needed some set of
> > policies or guidelines for handling memberships to the QA FAS group
> > for adding karma to the packages within the critical path of F13 (or
> > Fedora CURRENT_RELEASE+1). I volunteered to draft up such a document
> > in the wiki and I snagged a little bit of the wiki mark up from the
> > Ambassadors join page as a template, so thanks to who ever authored
> > that one.
> > 
> > Some notes on my Draft, I thought of putting together policies but I
> > don't entirely find this a policy style situation but I consider it a
> > "case by case" basis just as the Proven Packager process is. Its
> > essentially a "does this person do consistently good QA work?"
> > situation that (in my opinion) should be under review by peers to
> > decide their state of readiness to be responsible for karma that goes
> > into the Critical Path packages.
> > 
> > Ok, intro and disclaimer aside. Here's my proposal:
> >
> > 
> > There are some details on the mentors concept that I think would need
> > working out (denoted by the FIXME bit) that I assume can be worked on
> > at the next QA meeting.
> > 
> > Questions, comments, and snide remarks welcome!
> I think this is good, and it dawns on me that releng doesn't need to
> duplicate this.  I'm not sure where the thought first came from to have
> crit-path voting come from both releng or QA groups, when in reality it
> could just be a single group that has members who span different areas.
> It would likely make code easier on the update side too, and make
> finding somebody to karma up your update easier too.  So I'm going to
> suggest we kill the releng side of this and just go forward with net
> positive karma from the QA group, and those of us in releng that want to
> have our vote count can go through the process of getting into the QA
> group.
> Sound reasonable?

The only thing that worries me about is that it sorta compromises the
definition of the 'QA' group in FAS. It's a hostage to future fortune:
right now all the 'QA' group would mean is 'people who can approve
updates', so it'd be fine...but what if we decide in future we want to
use it for something else? We're stuck.

So far QA was planning to use this as an opportunity to actually define
a sane policy for the FAS group so that it really reflects the people
who are in QA, which would be nice to have for the future.

I think if we just want to make it one group, we should make it a *new*
group, called update_approvers or whatever. Members of QA could
automatically become members of that group, I guess.
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org

More information about the test mailing list