"What is the Fedora Project?"

Josh Boyer jwboyer at gmail.com
Thu Oct 15 23:45:44 UTC 2009


On Thu, Oct 15, 2009 at 05:11:33PM -0400, Máirín Duffy wrote:
>> Show me how we slow down without overlapping the RHEL case up there? B/c
>> I'm not real sure there's all that much room.
>
> * Focus on stabilizing a core of packages/applications, like folks are  
> saying, a critical path. RHEL focuses on stabilizing a much wider  
> breadth in comparison.

We're already doing this in rawhide.  I plan on trying to extend it to
stable releases as well.  I'm pretty sure we've said this in this thread
about 15 times now as a "should", when in reality it's already an "are".
(Not picking on you specifically, but we should really stop bringing this
up as something new to try.)

> * RHEL can not always carry the latest stable release of a given piece  
> of software, can it? Fedora should always carry the latest stable.  
> Fedora is released every 6 months. It's going to have newer versions  
> compared to RHEL no matter what; RHEL's major release cycle is much  

Newer versions does not equal 'latest stable'.  Something we should all
keep in mind: version numbers have a varied meaning (if they have a meaning
at all) across the FOSS world.  So if you're looking for 'latest stable'
you can't blindly assume newer is better.

(Yes, you probably know this.  But if we're trying to be clear, we need to
be crystal clear.)

> * Make it a priority to build tools in Fedora to make reporting the  
> instabilities that do make it through a complete breeze to do. Also work  

Already doing this with ABRT.

> on tools to make that stream of testing feedback dead easy for  
> maintainers to consume. Drive for a greatly-elevated amount of community  

Bodhi karma (+/-) is not dead easy to consume?  Or did you mean 'dead easy
for testers to submit'?  I think we can make progress on the latter, but
I'm not really sure how you can get much easier that the karma scheme we
have today.

Also, it doesn't matter how easy it is to consume if nobody bothers to
_give_ feedback.  So I think focus on increasing feedback is a better place
to start.

> don't. You pick a core to make sure are stable, but the very presence of  
> all the outliers differentiates Fedora as well.

You said the 'C' word.  You have no idea how hard we have tried to avoid
the 'C' word in all the discussions and proposals about critical path.

josh




More information about the fedora-advisory-board mailing list