On Mon, May 01, 2000 at 11:10:44AM +0200, Hartmut Koptein wrote:
> > Why must we go for shorter releases? Why is it so important? Why is a
> > release once a year or so not okay?
Release dates for comparison: (from the debian-announce archives)
version freeze-date release-date development/freeze
1.1 ? 1996/06/17
1.2 ? 1996/12/12 6 months
1.3 ? 1997/06/05 6 months
2.0 1998/02? 1998/07/23 8 + 6 = 14 months
2.1 1998/11/03 1999/03/10 4 + 4 = 8 months
potato 2000/01/16 2000/06? 10 + 4 = 14 months?
Hmmm. Clearly the best thing for our release process would be to get
Bruce back. And I may be hugely biassed and inaccurate in my worldview,
since I came in right around the hamm freeze.
Addressing your list, though.
> * people reject the use of 'unstable'
> * better and newer packages (bug-fixes), not so fast outdated packages
> * less dependency troubles
> * support for new technics (riserfs, ...)
> * continuous work, better update capability
`testing' should address all of these directly. Note that the point of
`testing' is that people who can't cope with packages more than a couple
of months old don't even have to consider running `stable'.
> * hopefully less RCB's
> * less bug possiblities, because of not so big package version updates
Note that these aren't particularly related to how long we've spent
actually working on the distribution. The slink and potato freeze times
are comparable in spite of the huge difference in how long we spent
developing them, for example.
> * we reflect that debian is a mainstream distribution
Windows is a mainstream operating system, and it releases every two or
three years or so.
Personally, I'd think that one interesting mainstream market for Debian
is in user desktops, which seem to me like they'd mainly want easy mass
installs that don't have to happen very often. debconf will hopefully
fixes the former, but the latter's hindered rather than helped by
frequent releases.
There's something of a contradiction here, though: automated mass install
via debconf isn't ready to roll with potato, so in one sense it'd be
quicker to achieve this if we fixed debconf and release ASAP. But then
again, I'm not convinced that taking an extra six months is going to
really hurt anyone.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG encrypted mail preferred.
``We reject: kings, presidents, and voting.
We believe in: rough consensus and working code.''
-- Dave Clark
Attachment:
pgp8P0ULMhJjl.pgp
Description: PGP signature