Response to "Getting Fedora Out of the If-Then Loop"

Mike McGrath mmcgrath at
Thu Feb 18 18:02:31 UTC 2010

On Thu, 18 Feb 2010, inode0 wrote:
> Try thinking about what I wrote ignoring a "target audience for the
> project" because that wasn't the target audience I was talking about.
> The default spin surely isn't just a pile of random bits. Those who
> create it have to have a target audience in mind even it isn't nicely
> stated on the wiki and well understood at a conscious level. So the
> question comes down to who is going to decide what that audience is.

Perhaps this is the disagreement.  As far as I can tell, the default spin
pretty much is a pile of random bits.  It gets decided by lots of people
(comps and the ks spin) with, AFAIK little coordination or discussion.
The only limit being a technical one because of the size we aim for.

Not only is it not stated on the wiki but I suspect it's not stated nor
understood by anyone in the process who makes the spin.  I'd hope I'm
wrong, if anyone involved in that process disagrees with me please correct

> Those who create the spin? The Fedora Project Board? As I stated in
> the following quoted hunk there are reasons for considering moving
> this decision to the board level and various parts of the project
> could benefit from doing so. And my concern is that doing this at the
> board level introduces a distortion into the upper parts of the
> pyramid. The consequences of introducing that distortion aren't known
> (life isn't that simple) and the board can conclude they are nothing
> to be concerned about given the benefits they expect in other areas if
> that is their judgment.

Why do it at the board level?  Because even when each team is picking what
packages should be on their spin.  They can only pick from a single rpm
version of each package.  Each package made by a packager for whatever
they think Fedora is.  When you install a package, it pulls in all sorts
of other things, those other things pulled in based on the best guess of
the packager.  Should X pull in the cluster suite or be broken into a sub
package?  What would the user prefer?  What user?

You know what?  I've explained this 1000 times at this point and you know
my point of view very well just as I know yours.  If you really care about
doing something about it (as I did).  Join the board.


More information about the advisory-board mailing list