Re: woody stable?? when?, measuring the notion of stable releases
On Sun, Nov 04, 2001 at 08:13:14AM -0500, Walter Tautz wrote:
> On Tue, 30 Oct 2001, Colin Watson wrote:
> > On Mon, Oct 29, 2001 at 06:53:56PM -0800, Karsten M. Self wrote:
> > > Incidentally, where is the release process / requirements documented?
> >
> > There are some vague guidelines in the developer's reference, but as far
> > as I know it's mainly documented in the oral history of the mailing
> > lists.
>
> And therein lies the main flaw of debian. The process of determining
> when a distribution is ready should be based on some kind of
> quantitative measure.
Debian's run by volunteers, so it tends not to go in for heavy-duty
management. Also, I wasn't necessarily suggesting that there is no
measure of when a release is ready, just that it's up to the release
manager, not carved in stone anywhere, and subject to change from
release to release.
> Also, once woody arrives on the scene as the next release a more
> vigorous revision process should be instituted via the pools mechanism
> which presumably will allow a more understandable upgrade cycle.
Many of us hope and work for more frequent releases. Testing didn't
really help for this release in that regard because it took a long time
to get it consistent and get the majority of active developers used to
it, but with any luck it should work better next time round.
If you're interested in this, please help (http://qa.debian.org/).
> i'd love to try to come up with a `measure' of stability..
As I said, it's usually very obvious when a release is ready.
http://bugs.debian.org/release-critical is a good start.
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: