I recently wrote a letter to eMachines. I have an eMachine at
home, and several of my colleagues have them at work. All of them
run Linux, and Linux only. My letter to eMachines discussed this
fact and encouraged them to consider pre-installing Linux for the
home desktop user. I pointed out the fact that they have experience
selling cut-rate computer solutions. They are currently selling
computers that push the lower limit of costs without requiring the
customer to sign up for a service. Their company description makes
it clear that they depend on selling packaged advertising as part
of their revenue stream. I pointed out that every machine they sell
that is converted to Linux is a loss to this model: they still have
to pay the MS tax, but they and their ad clients lose the revenue
stream they expected. They would do better to offer a Linux
installation (with or without the ads), save the MS tax, and
possibly maintain the revenue stream. It makes sense, there are
good arguments for them to pre-install Linux, and I thought my
letter was very convincing (at least to me!).
But later that day, I had a realization.
Why should we wait for OEMs to pre-install Linux? Why should we
wait for them to get the ball rolling? If our goal is world
domination, perhaps we should help the OEMs. We should
organize a project to create custom installations specifically
targeting machines currently on the market.
After all, Linux has been built on home-grown support for
hardware. We haven't been waiting for monitor vendors, or graphics
card vendors, or anyone else really to support Linux for their
hardware. We have been doing it ourselves. Well, the machine sold
by some OEM is just another unsupported device that we should
I propose a project that creates different installations for as
many machines as possible, in their factory configuration. There
would be two ISO images available: an installation disk, and a 'try
The installation disk would assume a factory configuration for a
specific machine. It would take over part of the hard drive and
install the Linux distribution, with all of the drivers needed for
the hardware known to be included with the specific model.
The 'try me' disk would boot into Linux from the CD, use a
folder on the windows partition to store preferences, and run a
sample distribution for users who want to try Linux but are scared
of the install.
The project would define a 'generic' installation, and all of
the specific machines would only have to differ from that
installation by a limited number of drivers and other hardware
choices. Anyone who has installed Linux on a machine could
potentially contribute the information needed to define the
This concept would help the desktop battle in two ways.
The first benefit is to reduce the hurdle needed to perform the
installation, especially for the new users, the faint-of-heart and
the lazy. Today if you are considering installing Linux, you need
to do a little research to make sure all of your components work.
If you want to buy a machine off the shelf, you have to look in a
number of different hardware compatibility lists, etc. But, with
the proposed model, you simply go to a site and find the make and
model of machine you want to buy, download the ISO and bingo!
The second benefit, which may help the cause more substantially
in the long run, is to make an installation that the OEM can even
consider offering as a pre-installed system.
Imagine you are eMachines (or any other OEM). You've been losing
installations to Linux for a while without realizing it, or at
least without knowing how many have been lost. You've contemplated
offering Linux pre-installed, but know that it would take some
investment that you can ill-afford to get the system tuned as
needed. Suddenly an ISO is being offered that is specifically
targeted at one of your models. How trivial it suddenly becomes to
consider pre-installing Linux on that machine. It would certainly
help make the case if the number of ISO downloads for each type of
machine are also listed on the download page. If you are Compaq or
eMachines and you notice that suddenly thousands of people are
downloading a Linux specifically targeting one of your machines,
you have a huge incentive to include that installation.
A project like this could be run either by one of the existing
distributions, targeting their own distribution, or it could be
done in a distribution independent way. The differences between
different hardware configurations is mostly independent of the
differences between the distributions. The downloads could include
variants for some of the bigger distributions. Imagine how easy
this could be: You want a machine and decide you really like the
Acer FP2-T800A, which is of course only offered with WinME. You go
to the download page, and grab the ISO
'Acer.FP2-T800A.redhat.install.iso'. Burn it to CD, pop it in your
machine, and away you go.
More and more OEMs are offering Linux installations for servers.
(I was surprised to see Linux prominently listed on the Acer home
page while doing my research!). They will eventually offer Linux
pre-installed. But in the meantime, we continue to moan about the
lack of pre-installed systems. Until OEMs actually offer desktop
Linux installations, it is up to us to do the job for them!
Interested in submitting a Community column for publication on
Linux Today? Contact the
editors with a brief summary of what you'd like to write about.
Not all proposals will be accepted, and we do reserve the right to
Some of the products that appear on this site are from companies from which QuinStreet receives compensation. This compensation may impact how and where products appear on this site including, for example, the order in which they appear. QuinStreet does not include all companies or all types of products available in the marketplace.