[nSLUG] Finding linux products and software

D G Teed donald.teed at gmail.com
Wed Apr 11 21:03:22 ADT 2007


On 4/11/07, Michael Gillie <mikegpc at gmail.com> wrote:
>
>
> I find that going all hell west and crooked trying to find linux products
> and related software is a real pain in the back side, and I knew that there
> was a multiple flavored linux server out there, by gosh and by golly i'd be
> the first to be johny on the spot.


I've been through a number of Linux distros looking for the perfect thing.
I've noticed that Ubuntu and Gentoo, roughly the same age, have lost some
users due to the bumps along the way for maintaining the 3 to 5 year old
systems
against some backwards-compatibility related issues.  It seems to be a
pattern: distros cross this medium age and have hard decisions to make
as things move forward.

It is pretty easy to make a distro that installs and works today.

It is hard to build on a distro, provide a forward path, provide backward
compatibility,
ease of system maintenance, etc.

I just upgraded my home server to Debian 4.0 yesterday.  It was done on a
running 3.1 server, upgrade in place.  I operate samba file service, postfix
with S.A and amavisd and clamav, apache, mysql, dozens of little sample
webapps like webcalendar, smf, xtide, and more I won't bore you with.  The
point is,
nothing was broken by the install.  I don't think other distributions
succeed in
this area.  Why?  Because they don't put the time into testing and resolving
the upgrade path for thousands of packages the way Debian does.

The founders of projects like Gentoo and Ubuntu sought to forge their
own distro because they were not satisfied or they were impatient
with the others.  As new projects, they could take any direction, free and
clear
of any legacy issues. But after a time, they too are struggling to invent
solutions to upgrade paths.

My point is, every new distro ends up at the same place - struggling with
what might seem
like human issues, but are really systematic issues which are not
necessarily easy.
Beyond solving the technical problems in the distro, the organization needs
to
develop methods and systems so these can be handled in a regular rather
than exceptional way.  Mature distros have crossed that point and have
developed stable processes and products, as well as solid user communities.

The solution is not to make your own distro, but to get involved with one
that comes closest to what you need, or provides the flexibility to get you
there.

I'm very keen on this concept, as I see dozens of things which Gentoo
(for example) did well (e.g. Guide and How-To documentation, as well as
XML based documentation system) and I would love to see some of that
talent pour into Debian.  It seems some Gentoo developers have jumped out of
that project, so perhaps one or two developers might come aboard other
distros.

--Donald


!DSPAM:461d774e229211939212641!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://nslug.ns.ca/mailman/private/nslug/attachments/20070411/019d9586/attachment-0002.html>


More information about the nSLUG mailing list