Re: branding debian releases
On Wed, Apr 14, 2004 at 12:19:39PM +0200, Pim Bliek | PingWings.nl wrote:
[...]
} I think the first question is of which user you want to attract. A good
} system admin knows what stable/testing/unstable means, but if you want to
} atract John Doe to run Debian as a desktop, we need to think a different
} way. If you need to attract sysadmins, the stable/testing/unstable naming
} schema is sufficient.
The idea of renaming the releases is coming up not because of marketing,
or attracting people. It is coming up because the current naming scheme
is causing misunderstanding among people who are trying Debian out. You
can argue that Debian is intended for sysadmins or experienced users (I
don't think Pim is arguing that, actually, but I expect that there are
those who would), but that's irrelevant.
The fact of the matter is that the mailing lists and IRC channels see
the same misunderstandings leading to the same questions (and sometimes
tirades) over and over. It is in the best interests of those of us who
want to provide the kind of community of tech support for which Debian
is famous to eliminate these misunderstandings so that our time can be
dedicated to solving other problems.
[...]
} Nowadays, if a John Doe comes up to me and asks me what distro to use, I
} must honestly say I will not tell him to go the Debian-way. Too
} complicated, and the stable distro is way too much out of date. I would
} suggest Knoppix instead for instance. Unstable is a no-go although it has
} proven stable to me, it does sometimes haves its quircks when upgrading
} and is thereby not suitable for John Doe.
}
} So in my opinion, Debian is not really ready for John Doe, except for when
} he has a nice cousin who knows Debian and can install a good unstable box
} for him (and maintain it) :). This approach works well for businesses,
} where they have sysadmins installing the systems for the John Does in the
} company. But for a home user, I will not suggest Debian.
I recently recommended Xandros to someone. Honestly, if a user is looking
for a computing environment that Just Works, I'll recommend a commercial
distribution (with tech support) every time. This has little to do with
the Debian naming scheme, however.
} Anyway, I think changing the naming scheme is not of real use at this
} moment. It will not help John Doe, and the sysadmins do not need it.
But remember, we aren't trying to help John Doe, exactly. We're trying
to help ourselves by making John Doe less confused.
} However, I can imagine you want to attract sysadmins coming from a
} different background (Windows f.i.!) willing to try the Linux-way. Would
} be sure nice if they give Debian a proper try. If you want these people to
} understand stable/testing/unstable you *could* think of different naming.
} However, I think a prominent FAQ document on the same pages as INSTALL
} docs and download locations on the Debian sites would be more helpfull.
People don't read FAQs, no matter how much we would like them to. It is
far better to avoid confusion in the first place.
} My suggestions for new names:
}
} Stable --> CURRENT_STABLE
} Testing --> ALMOST_STABLE
} Unstable --> NEW_NOT_PROVEN
[...]
Hm. Too long for my taste. People aren't going to bother typing
something that long in IRC. I'd say we want pithy but clear. How about:
stable ---> lowrisk
testing --> current
unstable -> earlyaccess
I can see an argument that testing should not be called current, since
CURRENT means something different in the BSD world. At the same time,
consider how the releases are actually used. The stable distribution is
most suitable for a) installation, and b) stable, high-security,
low-risk servers. The testing distribution is used for desktop machines
because it has current software, but is unlikely to break randomly (yes,
it still happens occasionally). The unstable distribution is used
package-by-package on many testing-based machines, and is also used by
savvy sysadmins who consider access to the latest software versions
worth the risk of broken packages.
That said, I think testing could be equally well named either
workstation or nearcurrent and achieve the same goals.
Just in case I didn't make it clear earlier, I consider this a
usability/documentation issue, not a marketing issue. The intended
result of any name change is a reduction in (repetitive) questions to
the mailing lists and IRC channels.
} Pim Bliek
--Greg
Reply to: