Re: Debian: abandon ship?
I just thought I'd add my $.02, or what ever is the appropriate monetary
conversion, in this inevitable discussion.
First, I am a user. I write in fortran by choice, c and c++, in the past
when I was teaching, by necessity. I am not a systems programmer so my
contribution to Debian would, at best, be an application or two. Currently,
I use Debian everywhere except on my Beowulf for reasons I will get to.
I'm writing because I like Debian a lot, I think it has been a remarkable
accomplishment, and I worry about its future general utility for me.
The reasons I like it are:
2. The remarkable support on this and other lists
3. The file organization
4. Its historic stability
I'll note in passing that the first reason has lost some edge now that
RH 7.3 comes with apt-get and that there are ports to older RH releases.
The reason I worry is currency. I am not really concerned about the roughly
5 week delay since May 1, but the nearly two year cycle between potato and
woody. I'm sure all developers are not only thinking about this but trying
to address this. Indeed, it may be that the infrastructure put in place
will address this wait. I have read some but not all of the messages in this
thread and, boy, there are some strong emotions. I hope that over time some
of the users comments and suggestions will be taken seriously.
Some developers have pointed out that woody is essentially
released. For my home computer this is true. However, if the security is not
ready then, for professional uses, where one sits behind a firewall and is
required to keep up with security, it will soon not be a viable choice. I say
soon because I am assuming, possibly incorrectly, that all security issues
have been addressed up to, say, May 1. This puts all of us about a month
behind. This is OK but not great. As time goes on, I will have to change.
What is even more problematic is that Woody is not itself up to date with
XF86. I have no choice but to move to 4.2 because my graphics card is not
supported in 4.1 . I have tried to use some debian packages from Columbia
University, but these have several broken pieces wrt the debian gnome packages.
I have compiled 4.2 from source and this works pretty well, although it leads
to problems with Star Office. I continue to tinker to get the system working
well enough. That is not what I do for a living. I'm supposed to do science.
I'm not trying to be a scold. I'm relating what it means to be a pretty
loyal Debian user at this point. I am also assuming that the Developer
Community actually wants Debian used everywhere, so, even though some of the
responses could have been on the 'Developer Knows Best' show, I think they
will take the user messages seriously. The Developer Community will have
to decide whether the delay was an infrastructure issue or whether this
cycle is indemic to a distribution with so many packages and so many
architectures. If the former is true (as I ardently hope) then this message
is so much hot air. If not, then, unless you want an ever-shrinking user base,
you will have to make some hard choices. It is true and noteworthy that there
are so many architectures supported by Debian. I sense an either-or mind set
about this. Either all of the architectures are released at the same time
or you scrap them. Why not have different release dates for different
architectures? As it stands, an architecture with a miniscule user base holds
up everyone else. Anthony Towns has already bitten the bullet and thrown
out a number of packages from woody because they were not ready. Why not take
a similarly pragmatic approach on architectures.
Another way to deal with this problem is to decide that there some crucial
packages (like, for instance XFree86) that have to be up to date in stable.
That is, you actually update stable to keep people from walking away because
their hardware isn't supported. That way you could take you time on some
big development issues and still keep a relatively well-served user community.
As I said above, if infrastructure was the problem, and the succeeding releases
are 6-8 months apart, you can throw these suggestions in the garbage. If is
isn't then if something like these two suggestions is not implemented you
could, at the least and of necessity, lose many workplace i86 machines.
On Thu, Jun 06, 2002 at 02:15:34PM +0200, Kerstin Hoef-Emden wrote:
> On Wed, 5 Jun 2002, Sean 'Shaleh' Perry wrote:
> > Debian is run by a few hundred programmers who do this for fun.
> > Not profit. Because we do this for fun we choose where to spend our
> > time. For some people the mips architecture and the required
> > hacking is fun. Others are constrained by the hardware available to
> > them (some of our developers had m68k only access).
> This one point, why Debian is different to other distris, it is not
> as i386-centered. In giving up its multiplatform support, it would loose
> one of its most attractive and fascinating aspects.
> > Cutting back to ia32 (x86) would help, but the cost is not worth it.
> > Besides, Debian is one of the few dists out there supporting
> > anything other than Sun and ia32.
> Without m68k I would possibly never have discovered Debian.
> > Removing those arches would leave out many of our users and
> > potential users.
> This is true for me, my "productive system" is not intel-compatible.
> > Maybe this means we lose some users to Red Hat (or SuSE or whoever)
> > and their 6 month cd releases. Everyone has to use what works for
> > them.
> At the last meeting of the Linux Workshop Cologne, we had more Debian
> than other users, although some people from this list claim Debian to be
> delayed. Seemingly there are more important things than just being
> up-to-date with the latest software packages.
> Dr. Kerstin Hoef-Emden Gyrhofstr. 15
> Universität zu Köln 50931 Köln
> Botanisches Institut Germany
> To UNSUBSCRIBE, email to email@example.com
> with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org
To UNSUBSCRIBE, email to email@example.com
with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org