Well you know, now when free os's are finally making some headway amongst users drastic measures must be taken to maintain its "exclusivity". So it looks to me like all of the most widely used distros aren't "the real thing" and therefore don't count.

I can just imagine what FSF meetings are like. Everyone brings their own desktop computer with them because after finally getting their laptops to hibernate they can't wake them up and everyone's tangled up in ethernet cable because no one can get their wireless cards to work.

But, on a more serious note, you can also argue that for a distribution like pure:dyne (as a media artist's distro) to attempt to adhere to the FSF's definition would very likely defeat some of the value of the distro. No?

Pall

On 2.4.2009, at 22:45, Aymeric Mansoux wrote:

Rob Myers said :
2009/4/2 Javier Garcia <[email protected]>:

there is a list here of linux distros that FSF recommends and qualified as
enterely free.

If puredyne is completely free it should be added, dont you think?

It would be great if pure:dyne was completely free. Whether it is or
not depends on which parts of Debian pure:dyne uses.

Well...

Looking at the list of officialy supported free OS, I wonder how they
do it, or if everything was carefully checked before they got listed.

For example, I suppose they have absolutely no non-free firmware?
so that means none of these distros have the firmware from the Alsa
project? or no non-free firmware for x or y hardware? Do all the
software they bundle is released under a free license? Does it take
other point of view on free software (re infamous DFSG vs GFDL) into
account? I also suppose they all use the linux-libre kernel? Or is the
kernel the exception for being listed as free OS? What about the
artwork, fonts, etc, do they provide free license information for every
single bits of non-software parts?

Also, I read:
http://www.gnu.org/philosophy/free-system-distribution-guidelines.html
"What would be unacceptable is for the documentation to give people
instructions for installing a nonfree program on the system, or mention
conveniences they might gain by doing so."

So breaking that rules means you are not providing a free OS I suppose.

That means none of these projects are actually explaining their users
in their documentation (is wiki documentation? mailing list? IRC logs?) how to make their graphic card work using binary blobs, nor explain how
to make their Linux supported hardware work when it relies on non-free
firmware or microcode? (which by looking at the libre deblob script...
are quite numerous!)

What happens when someone has, say an old ATI, that is now not supported anymore by the company, but has, thanks to AMD and X.Org, a free driver that provides a "good enough" support for most basic features? This free driver will not work fully anymore, because the microcode it needs, that
was provided so far by the kernel, is not not shipped in those free OS
(apparently) and there will be no documentation for the user on how to
use his old card with a free driver because it would not be a free OS
then?

What about the recycling of old machines that relies heavily on kernel
drivers (microcode and firmware included) because the hardware has been
long gone unsupported by their original producers? (when they still
exist).
Only a few old models will pass the free system test. What kind of
machines would be OK then? There are tons of left over and wasted
hardware lying everywhere that can be used to be good enough multimedia systems but I suppose some will never be able to run a truly free system
then? (unless free firmware and microcode get written which is not
necesseraly going to happen anytime soon).

Of course the FSF is listing recommended hardware for free systems, and
that's helping a lot when you're shopping (if you don't know this link
bookmark it, and contribute to it if you can).
http://www.fsf.org/resources/hw

... but at the same time, they can send mixed messages. While reading
through their hardware recommandations, I just see now:
http://www.fsf.org/resources/hw/systems

"As video card we recommend something made by ATI, because AMD (ATI's
parent company) has been releasing detailed specifications for a host of
ATI cards. This has made possible the development of a free software
driver that can do 3D as well as 2D. This driver is called 'radeon'
http://wiki.x.org/wiki/radeon, and it has 2D and 3D support for ATI
cards of the R100, R200, R300 and R400 series."

but these cards won't work completely because in a libre kernel the
microcode to make the acceleration possible has been removed...
http://www.fsfla.org/svn/fsfla/software/linux-libre/scripts/ deblob-2.6.29
... and there are still no free alternatives.

and making it work without seems rather difficult if not impossible
http://wiki.gnewsense.org/ForumMain/AQuestionForSomeoneWithARadeon7500

Similarly, there are computer dealers that are selling hardware with
gNewSense preinstalled, like: http://laclinux.com/ (I pick this one
because apparently they give some of the sales money back to the FSF).
But then how does such hardware work? For example the IWL firmware for
the wireless is gone from the libre kernel, so you get a half working
machine in the end? Is it a good demonstration of free software in the
end? or could it be that in the end everyone put some blobs back in
their system to make it work?.... of course all undocumented and in
secret... ;)

While I see the point the FSF is trying to make with this free system
guideline and while I absolutely respect completely everyone who is
actually trying to make this a reality beyond a simple "proof of
concept" (in the end it would save a lot of headache to have proper free
systems and the hardware that goes with it), I find it still very hard
to make it practical in real life situation...

and to conclude:
http://thunk.org/tytso/blog/2005/04/11/debian-the-fsf-the-gfdl-and- kernel-firmware/


a.




---
[email protected]
irc.goto10.org #pure:dyne


---
[email protected]
irc.goto10.org #pure:dyne

Reply via email to