Selling systems with Fedora preloaded.

Gain Paolo Mureddu gmureddu at
Mon Nov 28 06:50:44 UTC 2005

I really appreciate your comments guys, even though I may sound like a 
stubborn nonsensical guy at times, I'm trying my best to get this right.

Thus far I've gone through the nVidia and Flash licenses, they allow 
redistribution as long as the binary part of the packages is not changed 
(in the case of Flash, that's the package in itself and in the case of 
nVidia's driver, that's the X aspect of the package, as the kernel-side 
portion may require patches to get it built). In any case, both allow 
redistribution. I took a look at how other vendors are putting systems 
together, in particular HP workstations with some or another flavor of 
RHEL installed. We're not planning going with RHEL for a number or 
reasons (and thats not necesarily related to support subscriptions, more 
on that bellow). What I learned of how HP distributes their Workstations 
is by asking their users to log into their website to finish 
configuration of the Workstation (most likely to allow for third party 
software to be installed and properly acknowledge the users). So that 
there could be sort of a solution to our problem (details on this 
project and projected scope, etc, in a bit). Many of the extra packages 
we want to include are actually part of Fedora Extras, but that got out 
of Core (for some obvious and some not so obvious reasons), but which we 
consider could be considered as pluses, especially for our targetted 

The most spikey issue is of course that of Multimedia (as I have 
mentioned before), as people currenly expect to be able to have some 
sort of multimedia capabilities... Still while not directly providing 
the packages, would proper documentation on how to install them, plus 
disclaimers that if the user so chooses to install, say a DVD player, 
even though the validity of libdvdcss in Linux is doubtious (at the very 
least), the Fedora Foundation, the Fedora Project and ourselves (system 
builders) cannot be held responsible for the use given to such 
applications and tools, are not liable to responsibility from legal 
issues of any kind derived from the use there of? Sorry if my question 
is a bit confusing, but what I mean is that if we warn the users, even 
though providing the info on how to install such applications, but not 
directly providing the applications per se, and stating that any 
problems derived from the use of such apps, is the sole responsibility 
of the user and the user alone, would that still be a violation to the 
guidelines of Fedora Foundation and trademark use?

Rahul Sundaram wrote:

> Hi
>> Guess, Fedora is best suited for individual use only... As going
>> through all the restrictions, and balancing what most users expect to
>> find in their comptuers, it'd deffinitely be hard to market such
>> computers. Despite the computer's raw power. As I said earlier, what
>> worries me the most is the hardware part, as I can leave the system to
>> a default (kickstart) installation, letting users configure their
>> users, change root's password, etc., but (and I would too) users
>> expect the hardware they buy a new system with to flawlessly work with
>> the OS the system shipped.  This is what leaves me worried. As these
>> are the rough specs we thought of the systems:
> Like you have mentioned Kickstart has all sort of hooks for OEM to use 
> so the infrastructure to do more than individual deployments is 
> certainly there along with GFS, Xen and so on. Jesse Keating did a 
> presentation on Fedora for OEM distributions in FUDCon1 which you 
> might want to read

Thanks a lot, Rahul. I'll certainly take a look at it! Certainly 
Kickstart would be the way to go, plus taking advantage of the 
capability of Anaconda to install extra disks from the first boot 
interface. Whether the user decides or not to use the "extras" disk, 
would be up to him/her, and as such the Fedora intallation would be 
safeguarded that way, because up until that point, the installation will 
be a pristine Fedora default.

>> We've thought of a few ways to walk around this issue, like if
>> we just leave Fedora be and go for another distro (we wouldn't want to
>> do that, though) or offer the drivers as a separate disk with
>> installation instructions, and probably those packages we would have
>> had added to the system... BUT this could also in itself be an issue
>> if in anyway there's a restriction to do this as well. I'm going
>> through the licenses of Flash, RealPlayer and the nVidia (and ATi)
>> drivers as well... I didn't expect this to be easy...
> I cannot offer legal advise but here are some of my personal opinions. 
> Regardless of any distribution you use, you would have similar 
> trademark guidelines in place to prevent confusion.  As long as you 
> dont modify Fedora in anyway and simple redistribute it with the 
> additional packages clearly indicated as such the trademark guidelines 
> should not affect you. Do a license audit of the add on packages and 
> if the licenses allow redistribution without a EULA (Interactive 
> installations is against the design goals of RPM)  you can integrate 
> them within a repository and have a post installation hook to pull in 
> packages from a OEM  repository or design a custom application say in 
> GTK+  that has a druid  or even a simple shell script and zenity (part 
> of GNOME-utils) with fallbacks. The application would have the EULAs 
> which the user can agree to before getting the necessary packages 
> which can be launched on first login for the system 
> administrator/user. As long as you make it clear that this application 
> and whatever packages it uses is not part of Fedora this seems to be a 
> clean solution to me.

Well, just to round up a bit more what I said earlier. When looking at 
how HP configures their Workstations, a similar idea could be done here. 
Either require the user to visit certain webiste to gather additional 
information on how to set up their systems, or tell them up front in the 
documentation with a very visible, nice looking, EASY TO FOLLOW(1) 
installation guide and first steps with the new system, so that they 
understand that up to that point during first boot system setup, the 
system is a clean Fedora default installation, and what follows is our 
post-configuration to get all the additional programs and device drivers 
in place for intended system use (as advertised by us). The tricky part 
will be to have good Fedora advertising and still provide some extra 
functionality that will (hopefully) make users buy more systems from us. 
This blance between our intention to be true to Fedora, and yet have 
some value added to systems built by us, will be the REAL challenge. 
Especially to avoid striding too much away from a Fedora's default 

>> Just to clarify: Even changing default theme (to another GPL'ed one)
>> would cause an issue with the trademark? Even if the theme COMES with
>> Fedora in a default installation?
> All of the Free and open source software licenses allow you to copy, 
> modify and redistribute software licensed under them. Thats however 
> orthogonal to the trademark guidelines.
I'll take a look at that issue of Red Hat Mag. Anyway, I believe that 
the easiest way will be to have default settings an leave the users 
decide what they want... We like the clearlooks olive theme better... 
(though I have to admit I like it a lot too when combined clearlooks 
window border with Bluecurve Strawberry or Orange GTK colors)

> Legal like security is a field where it pays to be paranoidal. So we 
> have to assume worst case scenarios.  If suppose the trademark 
> protection guidelines allow the OEM to change the theme and if they 
> switch the distribution to use one of the al1y GPL'ed theme included 
> in Fedora as the default, that would be aesthetically non appealing 
> even while serving the functionality it is designed for, leaving users 
> of this modified distribution leaving a bad impression on what Fedora 
> is. So thats potentially a scenario that the guidelines are meant to 
> avoid. The alternative would be to get special exceptions which is a 
> hassle.

Yes, and because I know that is taht I'm trying to figure out what would 
be the best way to balance user requests with what can be provided with 
our systems, and having this wonderful distribution as the core of it all

> I would like to hear your plans with more details. How many systems 
> are you planning to redistribute Fedora?, market segment, timeframe etc.
> regards
> Rahul

The whole plan for the Fedora based computers is like this:

It all started when one of my best friends asked me to insall Linux on 
his PC and he became immediately hooked, switched in a matter of days...

According to what we have gathered thus far, there is a potential market 
in Mexico, quite big. Since we are a poor nation, and given the fact 
that usually computer hardware down here runs for at least twice as much 
as in the US, a cheapper alternative is needed. Not only that, but the 
systems sold by some of the big names in the industry can run for 
several thousand dollars for a top of the line system, while a mid-range 
system can still be quite pricey, we're talking that a Windows Media 
Center computer by HP can easily run for as much as $3500 USD. Mid-range 
and entry-level PCs while cheapper, usually lack a lot in the hardware 
department and even when they may have powerful components like 
processors and the like, usually the system is lacking in memory (like a 
P4 2.8 GHz with 256 Mb RAM with Windows XP on it) and are usually quite 
bloated in the software department (especially start up programs).

So we started  to think of ways to get better hardware at lower prices, 
and since we both use Linux we thought that it could only be natural to 
use Linux. But we've also identified some necessities from the users 
with whom we have most contact, like the need for really simple and to 
the point applications (we know Linux distros usually have plenty of 
those), easy enough to use interface and updates system (who can beat 
yum, anyway?!), etc. However, there is an increasing necessity for 
multimedia compliance, and here are the spikes. Because during the 
second half of the nineties, when multimedia systems bloomed, also 
bloomed the mp3 audio format, and very quickly the people started to 
compress all their CD libraries into their computers usin mp3... And 
when the world learnt that the format was not free and a license was 
required, that ruled out many Linux systems. At any rate, a lot of 
people has asked us if they'd be able to transfer their existin library 
into Linux, hence the need for a media player capable of playing mp3 
(and for some, wma, too).

The DVD issue is not as a big deal as the mp3 thing is, becuase a lot of 
people simply have a home DVD to watch their movies on, so they don't 
actually care about a DVD player, but they DO care about web and 
streaming video content... Another problem as the most widely used 
formats aren't free either, Microsoft's Windows Media Video and AVIs are 
quite common place on websites, not to mention other proprietary formats 
like Real Video and QuickTime. Was because of these "needs" that we 
decided that maybe including a media player like mplyaer, VLC or Xine 
could be a good idea, but when looked from above, it actually doesn't... 
Even though simply not providing a means to play this content could be a 
nay-say for many users.

Another issue we found that users were constatntly telling us, has got 
to do with security, virii and all those exploits of which Windows has 
been subject of as of late. Fortunately Linux is inherently safer than 
Windows, anyway, the real reason for that is that with all the spyware 
that some Windows computers mange to get, the performance of a computer 
starts to deteriorate to ridiculous point, so people have actualy asked 
us about "durability" of a Linux system, said another way "How much time 
until it start crawling instead of running". We believe Linux can help 
us there too.

So having this in mind and the fact that especially computer 
enthusiastic users have approached us becuase we are asidous Linux 
users, made us believe that we could sell what is still considered to be 
a higher-end system for quite a reasonable price, with support and a 
Free operating system installed.

So initially we'll try to sell the systems to computer savvy users, but 
the ultimate goal is to try and address as much as we can the needs of 
those "less literate" users. Having the hardware and the software in 
place is not all there is to it. Just like Apple did in late nineties, 
we believe that Linux should be sold "in style". That's why we set the 
hardware standards for our intended systems a bit too high and yet 
affordable. The sum of the parts of the hardware for a system like the 
one of my last post (and dpending on the graphics card the system has) 
can run for as much as $1200 USD (sacrificing in the graphics dept., 
with say a GeForce 6600 plain). Still a very imprssive system, yet quite 
affordable, as 1200 bucks is usually what an entry-level (with mid-range 
specs) costs. Of course there are systems for as low as 500 dollars, but 
they're too tight in hardware. The intention is to make the systems last 
(in the hardware dept., at least) for some at least 5 years.

For us profits would be in support and hardware sells... If we get a 
critical mass big enough, should be enough for the business to maintain 

However (and I did read about this in the guidelines), we'd have to 
figure out an advetisement campaing that not only puts Fedora on the 
radar of buyers, but also would not be in contraposition with said 
guidelines to not be misleading and stuff... and being this especially 
true if promoted as a gaming rig (for instance) with the waves of 
comercial games that have announced Linux support and all, as 
fundamental hardware support (in this case graphics) is not part of the 
distribution. Other market segment we'll try to get into is the small  
to medium office desktop, as we've had some possitive feedback in this 
regard. Corporate scale is not part of our scope, at least not yet. And 
for these systems, while still keeping much of the fancyness of the 
hardware, we've agreed not to soup up the graphics hardware, as such 
embedded Chrome chips are ideal in this scenario, as they even have 
kernel-level DRM drivers nowadays, so even graphics 3D accelartion is 
possible (if not as fast and fancy as with the "higher"-end systems we 
are planning.

As I said before, multimedia is a quite a big deal, even in an office 
environment, where the users may have setup a VoIP prgram, and they'd 
still want to be able to listen to their radio stations or music 
libraries, that's why we've also payed special attention to audio 
hardware, and deal with that which we know will be able to offer 
customers that without worrying about if they could dmix Skype or 
TeamSpeak while they're listeningto music, playing their games, watching 
videos or whatever. Right now we have very narrow options for this 
hardware mixing requirement (as I said my previous e-mail), and though 
there are more options, we've been unable to find (here) hardware based 
on other chips which are known to have good  ALSA hardware mixing 
support (like Trident's 4D Wave, besides the ones mentioned)

So pretty much these are as many details as we have right now... We are 
working our hearts out to research more about this legal mumbo-jumbo. We 
plan on developing these plans in the upcoming 6 months. The goal is to 
try and see how many systems are we able to get people interested on... 
I just hope we could find a balance between "cool" factor and legal 

So I guess that during the next 6 months our milestone will be set at 
approximately 1000 systems (conservative figures, I know).

(1) I've actually seen people render simple things into quite 
complicated matters.

More information about the marketing mailing list