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


Mikkel L. Ellertson wrote:
JD wrote:
I understand Mike Chalmers frustration with the release
frequency. No one is actually forced to re-install a new
release. The support cycle for each release extends to
about 18 months.
I am however in agreement with Mike's basic proposal: 
that given ANY fedora installation, there should be
an easy and seemless way to continually update all the
packages, and the kernel without having to re-install.
I think this is the general weakness of all linux distros.
While many will issue numerous reasons why this is not
possible, due to the domino effect of dependencies,
I think that the manner of how the dependencies are set up
needs to be altered so that none of the packages (including
the kernel) have any OS release version in the dependency
(such as foo-1.2.3.fc8....etc). The dependencies and the
packages should dump the 'fc??' (and here I am referring
not only to fedora but to all distros as well) and depend
solely on the package version and architecture.
Some may argue that this will wreak havoc with people
installing such packages on  non-fedora distros when
installed by inexpert users.
To that we need to say that even this dependency ought
to be erased. Let the system updater (such as yum) take
care and resolve the dependency issue. What I am driving
at is that there really OUGHT to be just one LINUX and all
packages out to be built for just ONE LINUX. Let the distros
distinguish themselves by their logos, initial installation
UI bells and whistles and the desktop backgrounds, even by
unique user tools. But the rest of the open source packages
from the topmost application to the kernel ought to be
for just ONE LINUX! 
Many people using linux today have absolutely no recollection
of why Microsoft eclipsed Unix. All unix vendors had their
proprietary Unixes, and they were not compatible. Major customers,
especially the government (where the big bucks come from) decided
to dump all these non-compatible unix systems in favor of windows
on every desktop and most server installations. Incompatibilty
and interoperability issues went out the window (pun not intended).
So, who are the big commercial Unix vendors today?
Arguably Sun Microsystems' Solaris, and IBM's AIX. But we all know
what has happenned to the installed base of Solaris and AIX.
It has shrunk drastically and continues to do so. Ever wondered
why IBM jumped on the Linux bandwagon?

I think Linux is repeating the mistakes of unix, and I can see
the writing on the wall. Linux will eventually kill itself and become
a mere novelty for a very minority number of worlwide users.
Microsoft need not complain nor worry about Linux contiuing to steal
customers. Microsoft COULD encourage divergent distros of Linux just
to speed up the process of creating all these incompatible distros
and it will win the battle. The Borg will indeed reign supreme.

Hello Linus - I  hope you wake up and smell the Borg :)

Are you a troll for Microsoft, or are you missing the hole point of
Linux? There should NOT be just one distribution of Linux. One of
the strong points of Linux is that it is not "One size fits all" -
you pick the distribution that comes closest to your needs, and then
fine tune it to your needs. This is one reason Linux is gaining
market share - you don't have to adapt to the OS - the OS adapts to you.

As far as distributions being incompatible with each other, for the
most part, this just isn't true. With the exception of some version
conflicts, they can all run the same software. There are packaging
differences between distributions, but even that is not that wide
spread. There are options to take care of that as well. There are
converters between packaging types. And if all else fails, you can
build your own package from the source. Where you can run into
problems is when you are back porting a package to a distribution
that is using older libraries. In that case, it usually works better
to start from the source instead of trying to shoehorn in binary

As far as UNIX losing its market share, it is mainly do to 2
factors. The first is that the market grew, and most of the growth
was in PCs that could not run the existing versions of UNIX. The
different UNIX distributions were designed to run on mainframes and
mini-computers. Until the 386 machines, the memory management was
not there in Intel CPUs. Now the different Linux distributions are
filling the need for UNIX on PCs. Linux fist made inroads in the
server market. It is expanding into other markets all the time. It
is interesting to see how many supercomputers are clusters of Linux
machines. It is also expanding into the small end of the spectrum
with embedded systems.

The most interesting part of all this is that Linux distributions do
not really have a marketing department. There are companies offering
services for people that want to deploy Linux. REdHat is probably
one of the best known of these. But there a lot of small shops that
will customize open source software to better meet the needs of your
company. It is a different model of making money off of software.

Mister mikkel,
So, the very first statement you made is calling me a troll for MS.
It is fitting that you call people names. It exposes your mental capacities.

I have been a unix user since the early 70's. Were you even  born then?

I have worked on the old unix version 7 kernel, the BSD kernel, the SVR4 kernel,
the Solaris Kernel and the AIX kernel (not necessarily in that order).
Perhaps at the time I was doing all this,
you were in.... kindergarten???? or grade school? or high school?

I saw the horrific problems companies were having with software packages that
would only run on some particular brand of unix, but not on others - so there
were multi brands of linux installations that required different people with
different expertise to update, manage and administer and secure.
At government installations, the picture was much much worse.
What made unix successful became it's own demise.
I think the same is happening to Linux. It's success is breeding incompatibilities
between the various distros, and that will kill it's possibility of taking the lead
from windows.
Real success should not be measured by how many different
candy flavors the kids can sample linux in. Real success is when more than 50% of
the private corps and all the gov offices  adopt just ONE LINUX. That's when MS will
really panic and resort to nasty behavior - similar to (but far nastier than)
your first statement in your reply to me.

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