Re: LCC and blobs

[Michael Poole <mdpoole@troilus.org>]

Goswin von Brederlow writes:

> Brian Nelson <pyro@debian.org> writes:
> 
> > You aren't reading what I've written.  Virtually 100% of firmware
> > out there (included on the device or loaded externally) is non-free.  By
> > your reasoning, the entire kernel should be moved to contrib since no
> > free hardware exists on which it can run.
> 
> Sure it runs on free hardware. On 100% free hardware. Take a pen, a
> paper and the boch source code and run your own linux on the pen+paper
> system. :)
> 
> Ok, it's a bit insane, but possible.
> 
> But let me say it again: "What matters is if the firmware itself is
> distributable at all and if it is DFSG-compliant."

You contradict yourself.  You can execute the firmware-loading driver
on pen and paper also; it operates just as well as the rest of the
kernel does.  Less trivially, imagine a device that speaks the same
protocol as the "problematic" device+firmware combination -- with the
distinction of not needing firmware.  Since the software can use that
device instead, the status of the firmware is irrelevant.

Otherwise, we must move clients and servers for network protocols into
contrib if the other end is not implemented by software in main: they
do not function properly without the other end.  Some examples would
be things that speak AIM, MSN, Yahoo! messenger, etc.  Since some
suggest that Linux kernel modules should be moved to contrib while the
rest of the kernel stays in main, it seems reasonable that AIM support
for gaim, naim, etc. should also be moved to contrib.

Michael Poole