Re: July 2010: RMS about the Hurd

[Arne Babenhauserheide]

On Monday 09 August 2010 11:43:27 Michal Suchanek wrote:
> On 3 August 2010
05:14, Quiliro Ordóñez <quiliro@gmail.com> wrote:
> >> | Finishing the HURD
would not advance us at all in supporting these
> >> |   devices. The work
that is needed is at the driver and firmware
> >> | level. That's why our
high priority task list includes items relating
> >> | to free drivers, but
not the HURD.
> > 
> > Do you think that having a good Hurd would remove or
at least lessen the
> > problem of the non-libre firmware?

> The option to
reverse-engineer non-free drivers is a workaround but
> did not provide a
long term solid solution so far.

I think the long term solution is easily
stated: 

1. Don’t include unfree drivers in distributions (or at least
require the users to explicitely state that they want to give up freedom –
in Gentoo I have to add an entry to my /etc/make.conf for every unfree
license I want/need to accept).

2. Become so big that the hardware
manufacturers lose a major share of their money by not producing a free
driver. 

3. Provide a brand: Free Software compatible. Hardware which is
compatible to the major free systems (regardless of by virtue of the
manufacturer or because the firmware was reverse engineered) can get a cool
brand. 


I think point 1 won’t change with the Hurd. Currently it’s up to
the admin and the user can’t change the fact that he uses the unfree driver.
With the Hurd there could be a thinner free glue layer used in all drivers
and the user could simply decide to replace the unfree driver with a free
one. 

Due to easier low-level experimenting it might become much easier to
write new drivers, though, increasing the number of people who reverse
engineer for fun. 


For point 2 we’re on a good way. Statistic evaluation
(in german): 
→
http://draketo.de/licht/politik/leserbriefe/gnulinux-gewinnt-marktanteile

T
his started as an answer to a journalist who was reading the data
wrong…
Short version: 

* August 2008: 8,31 GNU/Linux users per 1000 Windows
or MacOSX users. 
* July 2010: 9,65 per 1000. 

Net gain of 16% in two
years. But mind trusting the data: It’s just browser statistics. 


Point 3
I don’t know. Is already someone working on a brand which manufacturers can
use to say Look here, this device is cool! It respects your freedom! Buy
it!“ – maybe allowing them to use the GPLv3 brand if all necessary drivers
of the device are under the GPLv3 (or compatible – they can just add some
wrapper code and the derivative work becomes GPLv3)? 

AdBard is a nice step
into that direction. 

Best wishes, 
Arne

signature.asc
Description: This is a digitally signed message part.