Draft: simple update description guidelines

[Jeff Spaleta]

On Tue, Jan 27, 2009 at 10:49 AM, Thorsten Leemhuis
<fedora at leemhuis.info> wrote:
> Well, we have done fine without special rules for update text many years now
> afaics. So why do we need them now? Why can't "best practises" do, if it
> worked fine for so many years?

Has it worked fine?  I've no way to really measure whether the
introduction of any optional "best practice" has resulted in a more
consistent packaging landscape. Do the documentation of best practises
as non-mandatory suggestions on workflow materially help the
consistency of the packaging quality?  I don't know, I've seen no
supporting evidence either way.

Historically, guidance has evolved to addressed identified
deficiencies, has it not? We don't really know if this is the best
approach, but to suggest things have been working is to discount every
single complaint that has resulted in a guidance change.  I don't
think you are making a self-consistent argument.   The fact that
guidance is evolving is a reaction to identified deficiencies.  What I
don't know really is if we've gotten to the point with guidance that
we are chasing down problems that are in the random noise of human
error.
In this case are we talking about guidance to address 10% of the updates? 30%?
How widespread does a problem need to be before we make packager wide guidance?
I don't know.

>
> Whatever:  I don't care much about this specific set of rules. It just seems
> to me a general trend that more and more rules and guidelines are put into
> place in Fedora-land; and more and more work is outsourced onto the
> packagers. Sure, some of the rules and the outsourcing is needed -- but I
> find this trend very alarming.

You'd probably make more sense to me if you were arguing about a set
of rules you do care about.  I would appreciate it if you would limit
using the word trend to describe something that can be depicted
quantitatively on a graph.  I would love to be able to trend the
impact of best practise documentation and guidance, but we don't
really have any quantifiable metrics here at all.

But here's something we can graph.  How has our contributor growth
month-to-month  release-cycle to release cycle looked like?  I think
we could probably did that out of fas.
Is our rate of contributor growth slowing?

-jef