[CentOS] EL 6 rollout strategies? (Scientific Linux)

[Gordon Messmer]

On 05/15/2011 06:10 PM, Johnny Hughes wrote:
> Where is Ubuntu telling people exactly where they stand on producing a
> their new releases.
>
> What about Red Hat ... how about Fedora.

I don't know about Ubuntu, I don't use it.

Fedora, on the other hand publishes their schedule:
http://fedoraproject.org/wiki/Releases/15/Schedule
And the release life cycle:
http://fedoraproject.org/wiki/Fedora_Release_Life_Cycle
And their release criteria:
http://fedoraproject.org/wiki/QA/ReleaseCriteria
And release engineering documentation, including the names of 
responsible persons and directions for getting involved:
http://fedoraproject.org/wiki/ReleaseEngineering
And standard operating procedures:
http://fedoraproject.org/wiki/ReleaseEngineering/SOP

The release criteria includes a Bugzilla list for a release blocker bug 
which shows users what issues currently need to be resolved before the 
release.  Users are very well informed about the state of the project.

Fedora uses Koji to build packages.  Users can view build logs in the 
Koji interface as well.

After building packages, maintainers push to Bodhi, where users can test 
the package and indicate success or failure before the package is 
finally published.

If CentOS were run even remotely like Fedora, these discussions wouldn't 
come up.

> Is there someplace I do not know about where these distributions tell
> you what they are having trouble building?

Apparently.

> Show me another list where the developers interact with the users as
> much as this one.

My interactions with Fedora's developers and maintainers have always 
been both pleasant and productive.

> CentOS has never been secretive.  We published examples of our build
> scripts for the RPMs and the disros, the mock we use and plague.
> Something Red Hat has never done.

It doesn't always seem that way to users.  Certainly, the trend has been 
to greater openness and more insight.  That has been encouraging.

In February of '09, Karanbir published a blog on the r-v-m routine.  I 
vaguely recall that sometime in the years before that he stated that the 
scripts used to build the release would not be released, which was a 
significant part of the reason that I, personally, have regarded the 
project as somewhat secretive.  More generally, I would describe the 
project as somewhat secretive by virtue of the lack of communication 
with users.  I don't intend to imply that the developers are malicious, 
just that many users clearly feel like they do not and cannot understand 
the state of the project.

Look, I appreciate the new QA site.  It's great.  I sort of remember 
someone linking to a page with a list of the tasks blocking the release 
of C6, even though I can't find it now.  That also makes me a happier 
user.  However, I can appreciate CentOS and the work of its developers 
without thinking that it is perfect, right?  Is that too much to ask?

> Can't you ungrateful bastards take the free software I make by following
> the licensing requirements and be happy with that?

Wow. I guess not.

All of that is more or less a distraction from the point at which this 
branch of the thread began.  One person suggested that 6.1 might take 
only a month, and that seems highly questionable.  Without making any 
value judgments about whether or not the distribution *should* be 
available in one month, or whether some other project can do it faster, 
and without questioning the competence of anyone, I still think it's 
legitimate to express doubts that a release can be made ready in that 
time frame.  There is no recent evidence that users can expect that.