Fedora 12 Features Proposed for Removal

Adam Williamson awilliam at redhat.com
Sun Aug 9 07:03:37 UTC 2009


On Sat, 2009-08-08 at 11:07 -0400, Josh Boyer wrote:
> On Sat, Aug 08, 2009 at 06:47:55AM -0700, Adam Williamson wrote:
> >On Sat, 2009-08-08 at 06:55 -0400, Josh Boyer wrote:
> >
> >> We have a Feature process now, and while I'm not a huge fan of process,
> >> this one is run very well by the Feature Wrangler.  It treats them all the
> >> same, they have to all follow the same criteria regardless of how 'obvious'
> >> some of the decisions are.
> >
> >...except I pointed out that we do version bumps of everything in the
> >distro from one release to the next (more or less), but only a couple of
> >them are flagged up as features. To which it was replied that this was
> >primarily for PR purposes. To which I objected that it was silly to
> >attach this kind of bureaucracy to something which was only happening
> >for PR reasons.
> 
> What does any of that have to do with the email I replied to?  You said it
> was silly to have a meeting to follow through with the process when everyone
> knew what the result would be, and I replied why we do that.  Now you circle
> back on some other thing?
> 
> I'm pretty sure you're just arguing for arguing's sake now so I'll just bow
> out and stop wasting my time.

Sorry, I just read it back to myself and realized I probably didn't
quote the right bit of your email, and my thought process wasn't exactly
clear.

It goes like this: there's consistency in terms of how features are
handled as part of the feature process, but there's no consistency in
terms of what the proposed features actually are or why they're being
proposed as features, so the process as a whole is not consistent and
produces this kind of absurd result. It's good that the feature review
process is consistent, and I think the process is well-designed and
well-run as long as it's being applied to things that sensibly
constitute features, but when it's applied to things which don't
sensibly constitute features for anything but PR purposes, it produces
silliness like happened this week (through no fault of the feature
process or the feature wrangler).

In short: just because there's a clear process that's consistently
applied, doesn't mean there's no problem.

To elaborate again (because I can't make anything short ;>), the flip
side is the process doesn't solve the problem you mentioned in the bit
of the mail I cut out - transparency and accountability - because we
_don't_ apply it to everything. The kernel version bump, for instance,
wasn't submitted as a feature. So if there were some problem with that,
presumably FESco would still be using a non-definable and
non-transparent process to make a decision...

-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net




More information about the fedora-devel-list mailing list