bugzilla triage madness :-/

[Les Mikesell]

Stephen John Smoogen wrote:
> 
> Well that comes down to managing expectations. Fedora is not a stable
> product in the sense that a washing machine is. It is a fastly moving
> prototype product using volunteer labour. Packages can be updated on a
> released version that has bugs that weren't in the 'tested' version.

Agreed, as much as I like to grouse about bugs myself, I do understand 
that _some_ of them would never be found and fixed if you didn't push 
them out to unwilling fedora users.  But as far as managing expectations 
  goes, I don't see anything on http://fedoraproject.org/ or the 
Overview page that explicitly says to expect bugs.  We may understand 
that 'latest in free software' implies that it is still in testing, but 
it is not really obvious.  I also don't see anything on those 
introductory pages that would encourage people to participate in rawhide 
testing if they really want the latest software.

>>  If the previous release was in fact "good enough", don't break the fedora
>> released version by pushing out new bugs that are known but not fixed in
>> rawhide.  A judgment call, of course.
>>
> 
> Every released package is good enough for some segment of the
> audience.

Yes, but there's no way for the segments to sort themselves out once you 
   push to the production release.  Someone who wants the newest, 
untested versions can always install from rawhide.

> And every product is going to have bugs. Some subtle and some obvious
> but non-fixable without adding more bugs.

Agreed, but some are showstoppers, some aren't.  How do you draw the line?

> If Fedora were a business with paid employees, I would probably look
> at reworking how bugs etc are dealt with:

And I'd wager that you would also try to enforce a much stricter policy 
on new, known bugs being pushed to customers as released products to 
hold the cost of supporting them down.  Since you don't have the paid 
employees to fix these after the fact, isn't it even more important to 
control them at the source and distinguish the willing rawhide testers 
from the users expecting the fedora release/updates to be a more stable 
product?  If you are looking for more and better quality participation 
on the BZ reporter side, I would expect the best way to get it would be 
to court rawhide testers or force people there by holding back on what 
goes into fedora releases until you have reason to think that you won't 
be increasing the released bugs.

-- 
    Les Mikesell
     lesmikesell at gmail.com