[fab] discussion topics for red hat ceo

[Stephen John Smoogen]

On 5/15/06, Michael Tiemann <tiemann at redhat.com> wrote:
> On Sun, 2006-05-14 at 16:20 -0400, Max Spevack wrote:
> > Howdy folks,
> >
> > The Fedora Project Board meets next Tuesday (5/16), and Matthew Szulik
> > (Red Hat CEO) is scheduled to be in on the meeting.
> >
> > I've asked the board members to think about some of the topics they'd like
> > to discuss with Matthew, but I wanted to throw it out to this list in
> > full.  What are some of the things you'd all like to see discussed?
>
> My question would be: If Fedora could do one thing perfectly, what would
> it be?  I pose that question to the FAB--how would you justify that
> specific achievable goal as the highest priority to Matthew, and I'd
> pose that question to Matthew--what does he think is the one thing
> Fedora should deliver if it could deliver one thing perfectly.
>

Hi Tim, I am a little confused by the question. One perfection is
pretty much impossible, and two every engineer wants everything to be
perfect. That usually causes a lot more headaches as whenever the word
"perfect" comes up people expect/push for it to happen. It also seems
to be the deeper reason for people's disgruntled-ness if they do not
achieve it. Engineers seem to be highly disgruntled if things are
99.997% there, and less so if they are only 60% there.

In trying to think of one thing that I would like perfectly.. I ended
up realizing that there were many things that were outside of
perfection.
  1) A regularly released distribution with the latest tools is tied
in with every outside project (GNOME, KDE, ALSA, kernel, etc) and with
show schedules conferences etc for developers.
  2) Cleanly resolving and closing all bugs of the previous release is
tied also with outside projects as many of the bugs that are seen go
upstream.
  3) Providing a perfectly documented system relies on the above two
questions (what is in the release, and what bugs were fixed or noted.)
  etc.

I am probably overanalyzing the question. I can see a list of things
that RH-Fedora does well at 68%, and some stuff at 95%, and maybe a
little at 99%. My view has been that Fedora should focus on the 80-20
rule per release. If 80% of the population is happy with the release
and the 20 % are being looked at in the next release, we should be
happy.

What I would like to have happen from the board and management (though
they may be outside of the control of both).
1) Not have the project leader of the month... as seemed to be the
outside view of the project for the last ~2 years.
2) Some official listing of 'long-term' (2-3 years) support for
Fedora. Red Hat has a good history of thinking on its feet and looking
at projects, acquisitions, etc and going "what the f* were we
thinking?" and closing them down before they became a long term
liability. This has been a good thing for the long term viability of
Red Hat, but has a bad side effect in that people on the outside
'selling' Fedora have to deal with the "Well how long is Red Hat going
to support this thing, this time?"

-- 
Stephen J Smoogen.
CSIRT/Linux System Administrator