Re: Social Contract GR's Affect on sarge

[Nathanael Nerode <neroden@twcny.rr.com>]

Michael Poole wrote:

> I want to distinguish between software and other data because I prefer
> to use English in a precise way,

If you want to do that, please use "computer programs" for the usage you're
thinking of.  "Software" has a precise, if hard-to-define, meaning, for
which there is no synonym, and using it as a synonym for 'programs' makes
it harder to use English in a precise way.  I've mentioned this a gazillion
times.

> and because I think that is
> consistent with the broader usage[1].
> [1]- See, for example, http://www.gnu.org/philosophy/free-doc.html

If you wish to distinguish between the freedoms desired for programs, those
desired for data, and those desired for documentation, you have to come up
with an actual specific distinction.

On debian-legal, people tried to come up with these for years.  Every
distinction between documentation and programs was refuted by someone who
said "I need that freedom for documentation, and here's why!".  Similarly
for standards.  Similarly for star charts.  Similarly for images. 
Similarly for sound files.  And so on.

Furthermore, if you want to distinguish between programs and other data, you
have to come up with a distinction which works.  It's been pointed out that
a 'program' is a very slippery thing to define and that most types of
'data' on a computer are in some sense programs, if you accept that perl
scripts are programs.  Perhaps one can come up with such a distinction, but
if there's no difference in the desired freedoms, what's the point?

Anyway, very few people seriously think that the "firmware" under discussion
isn't programs.

> That does not mean that software freedom should be the only freedom
> that Debian pursues, but it does not help to pretend that Free
> Software is the same thing as Free License Texts or Free Reference
> Documentation or Free Speech.
See above.  It doesn't help to say that they're different without explaining
*why* they're different.

>> I'd desperately like to see even
>> one reason that isn't "Think of all the other KeWL stuff we could put in
>> Debian".

<snip>
> People have identified specific tangible benefits from having binary
> firmware in main: primarily needing less effort by Debian to support
> certain hardware.

Unfortunately, that benefit is much the same as the benefit of putting the
NVIDIA binary-only drivers in main.  Since those are excluded, you need to
make a solid distinction between the two in order to be saying more than
"look at all the cool stuff we could have".  The distinctions made so far
have been attacked on many grounds, ranging from being vague and difficult
to apply, to being completely false distinctions.

A distinction given by Marco D'Itri -- programs running on the main CPU
versus programs running on a peripheral CPU -- is pretty good, but it seems
to be a dangerously arbitrary limit on freedom.   It's also hard to come up
with a decent distinction what's a peripheral and what's the main system,
what with the graphics processors often being more powerful than the
Pentiums.  And is the 80387 a peripheral?

>  Debian has decided those benefits are not worth the
> alleged cost in freedom of shipping binary blobs, but people can (and
> currently do) debate the value of the official position.
Yeah.  I just wish they would not rehash the same discredited arguments over
and over again.  Presumably it's because you haven't seen the prior
arguments, so I politely presented some of the highlights above.

-- 
There are none so blind as those who will not see.