Delays in package processing

[Jeff Spaleta]

On Dec 21, 2007 7:25 AM, Bill Nottingham <notting at redhat.com> wrote:
> Well, it depends. For the stuff I maintain, I tend to follow different
> policies depending on the package:
>
> - Library versions do not get upgraded in a stable release
> - The more core a package is, it only gets updated with occasional
>   severe bugfixes - it is not ever rebased.
> - More peripheral things, such as some desktop apps, get updated to
>   the latest minor bugfix version
>
> Hard to say how much this works for everyone, though.

Let's start with this methodology. Can we split the repository binary
packages into rough
concentric rings that embodies "coreness." Call it Spaleta's 9 circles
of Fedora Packaging Hell{tm} if you want.  And then produce some
aggregate stats as to how updates for those groups have been handled
for F7 and so far for F8.  Is the aggregate update policy inline which
the methodology you just outlined...update churn goes down as you move
into inner/lower packaging layers? Or is there an obvious deviation in
the aggregate for a specific layer?

-jef"and again by 'we' I mean 'not me'"spaleta