[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian Critical Mass


I'm a Debian user. So what I want to tell you
is that I have never seen Debian as a distribution but as a package
pool. To me Debian is a package format with some org thats maintaining
these packages.
Recently I had to upgrade to unstable/frozen because I needed glibc 2.1.
And you can believe I did. Using a distribution that cares for
stability, safety and consistency makes me feel good. In fact it is a
great feeling in the morning, when I get up. But it is pain in the ass
after the breakfast, if your stable and consistent operating system
turns out to be useless because it is totally outdated.
The problem thereby is not a too long release cycle. Other OSes have
much longer release cycles. One problem is that the software, that you
want to distribute, develops so incredable fast. An other problem is
that the number of software which you want to distribute increases too.
There are for instance more than 5 dvi viewers in your distribution.
Most of them are more than 5 years old (and since then not updated),
have a terrible ui and lack basic features. All of them appear again and
again in new releases. On the other hand I can not find up to date GNOME
packages for the stable release which I regard as very important.

Splitting up main into sections would only make sence if these sections
have clear dependencies. i.e. if you had 2 sections then only one may
depend on the other. Otherwise independent releases were not possible.
So you could make an uptodate GNOME release for an already stable
release, without updating the whole release.

Next you should decide which packages you really need in a release and
leave unimportant or outdated software out. I don't say that you should
stop supporting some software. These left out packages should still be
maintained and available but they should not hold back releases.
Actually none of the dvi viewer which I found were really usable for me.

At last I think, that you should try to find out what kind of users you
really want to serve. Remember what happend to whose who try to be good
everywhere! They end up being good nowhere. 
Corel is a Distribution. They have their target customers and tailor
their distribution for these customers. Debian is not a distibution.
Debian is a lot of packages with some installation tools. A package
pool. I know, that you cannot turn debian into a server OS or a desktop
So the only reasonable suggestion I have is splitting Debian into
several distributions which all use the same package pool (beat me if
you want). The idea of sections would be very helpful for this. The
advantage is that a webserver doesn't hold back a desktop system and a
server OS doesn't need the latest gui features. And users don't need 5
outdated dvi viewers ! 
All these Debian-dists would stay compatible by using the same base
system and the same packages. Further kooperation and koordination could
produce complete Debian releases. 

I heard some ideas about incorporating information about which sections
a package belongs and introducing a metastate between stable and
unstable where packages get in automatically after some time unstable to
assure that the most terrible bugs are out. In my opinion these are
first steps into the right direction.

So I'm just a user and don't know very much about the internals of
Debian management but these ideas would solve some problems for users.
However, I think the current problems will only become worse when the
number of packages increases. Unlike Microsoft which is the only company
that releases Windows OS there are dozens of distributions that
distribute the same software as you. For Microsoft its no problem to
wait another 6 months for a release. But if you don't manage to release
faster, then you will eventually loose most of your users to other


PS How many people do you think use stable Debian exclusive. I don't
know anyone.

Reply to: