Le samedi 08 décembre 2012 à 05:12 -0800, Dan Mashal a écrit :
This IS a rant. And this includes a few analogies. Some good, some
This is one of the reasons why I chose to run for board.
Nobody really knows where Fedora is going. It's like a too many chefs problem.
Sometimes Fedora just feels like a bunch of people/SIGs working independent of
each other (Red Hat included) that eventually come together to make the entire distro.
That's kinda how free software work, in fact, individual group of people
producing a small part of the system.
There needs to be a more concerted to have a direction. There needs
to be more
communication with end users and what they want from Fedora.
For example, the same thing happened with Gnome 3 upstream where a lot of developers
left the project due to a lack of a real vision or direction.
That's not really what seems to be seen with the graph at ohloh :
on 5 years, the number of contributers seems quite constant, and I think
the few people that left could be attributed to others external factors
such as :
* canonical focusing on unity,
* nokia's fate and the move on QT for meegoo
* sun/oracle merger and the various internal changes this could have
* problem at opensimus
), and similar likely due to the financial crisis or IT cost reductions
Companies are where you find lots of contributions, not end users.
In fact, I never heard anyone complaining about "kde is dying" while the
numbers are much more worrisome :
Maybe that's caused by some cleaning of their svn, or maybe that's just
nokia's problem ( and the huge layoff that followed ).
I would also add that if the switch to gnome 3 made enough people leave
the project, they would have gone to mate, and afaik, no one coding on
mate has a @gnome.org email. In fact, mate do take a lot of commits from
( ie, 4 out of the 5 commits merge on 06/12/2012 come from gnome )
People told "users would migrate to xfce", yet it didn't translate into
more contributors for the project :
( and you can check also for lxde, and others ).
So your assertion of "lots of developers left because of the lack of
vision" is IMHO wrong.
And I would not say that GNOME do not have a vision, quite the contrary,
they do have a idea of what they want to achieve ( at least, Jon McCann,
), and that's something that bother part of the user base, because they do not share
the vision. Google for "Design Principles for the Next
Generation Desktop" to see a talk about that, or look at the various
GUADEC talk and proposal in the past.
"Let's just include the latest greatest things that are cool
in the Linux world."
(FIRST on FEATURES) is part of the problem.
"Let's make it look like Ubuntu because Ubuntu is popular" is another
I have seen and I have a problem with.
In addition, LTS won't solve this problem. It goes against everything Fedora stands
So "first on feature" is a problem, but that's what Fedora stand for.
Yet, LTS would be bad because it goes against what Fedora stand for.
I fail to follow your line of reasoning here as it seems to contradict
A really bad analogy and half joking here: LTS should just be renamed
"Slackware". If you
get the analogy kudos to you.
In my opinion the vision needs to be changed. It feels like Fedora has turned into
Rawhide more than Fedora with 17 and even more so with 18.
You mean like people who are pushing features
) directly on all
stable releases ( https://admin.fedoraproject.org/updates/mate-desktop
), despites being frowned upon by the policy :
was part of the vision that the board proposed :
If you want to avoid Fedora 17 turning into rawhide ( ie, a rolling
release where is pushed latest version of packages ), I would propose to
start by following your own ideas, and help others to follow the rules
about it, instead of asking for others to do one thing, and yet do
another one yourself.
If there was more testing done with Rawhide then I wouldn't feel
like the 6 month
release cycle may be a bit too aggressive right now.
I mean the proof is in the pudding. Spherical Cow is almost 3 months late.
I see 2 months and 2 days, hardly 3 months late. Not to mention that
IMHO, part of the problem is due to christmas holidays, and
thanksgiving. ( and that would remove 3 weeks out of the 8 weeks late ).
So that's more 5 weeks, instead of 12 ( 3x4 ).
Anyway, your concrete proposal is ( cause this has been discussed to
death already, so let's go forward and fix the issue if this is
important ) ?
in fact speaking of more testing in rawhide, do you run rawhide ?
If not, maybe that's something that is part of the problem, and the
first step to a solution. IE, someone should volunteer, and get enough
feedback on why people in the fedora community do not run rawhide. Once
the problems are identified ( even roughly ), then the next step to
correct them could be found.
If the issue "people tell me to not do it", then we should change the
message. If the issue is "I did it but this $class_of_bug made me lose
months of work", then we should find a way to prevent $class_of_bug, and
make it know. And so on.
But just saying "more tests should be done on rawhide" doesn't make them
happen. You cannot force people to run rawhide.
Going forward, I would like to look at what the real vision and
direction of Fedora should be.
In my opinion it should be a distribution that offers the latest software, but at the
time keeping stability and compatibility a #1 priority.
Fedora community do not develop most softwares we ship, that's usually
the job of upstream communities ( such as gnome, xorg, gcc ). So you
should engage them into buying your vision about stability and see what
they need, and help them.
For example, I am sure the people who work on
to have their tools integrated into upstream projects, so warnings can
be emitted when ABI is changed, and pushing that as best practice ( like
pushing fuzzing or valgrind memory check ) would be a good step into
Or maybe add this into autoQA ( and fix
At the end of the day the people that make a distro popular are the
users, developers and sysadmins.
Sysadmins want things to just work as they have been for the last 15-20 years with some
that don't require them to go back and relearn everything. They don't have time
I do not know for you, but as a sysadmin, I do not care having to learn
new things. For example, I am looking forward cleaning some cruft with
systemd migration. I do not expect to live in a bubble unaffected by the
fact that the rest of the world is moving. 15/20 years ago, we didn't
have cloud/virtualisation, BYOD trend, hyper connected work force, etc.
So sysadmins and IT has to adapt daily. And changes into systems are
just part of the job. People are paid because they are able to adapt.