[directory-discuss] s/w that requires a middleman to liberate it -- is it free? (was: GNU R situation)

[Anonymous]

Svetlana Tkachenko said:

> You may get a copy at
> http://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar[1].gz[http://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar%5B1%5D.gz],
> on my website which hopefully does not run non-free scripts.

===8<------------------------------
http_proxy=127.0.0.1:8118 wget 
'http://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar[1].gz[http://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar%5B1%5D.gz]'
Warning: wildcards not supported in HTTP.
--2017-02-03 19:16:25--  
http://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar%5B1%5D.gz%5Bhttp://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar%5B1%5D.gz]
Resolving svetlana.nfshost.com (svetlana.nfshost.com)... 208.94.118.221
Connecting to svetlana.nfshost.com (svetlana.nfshost.com)|208.94.118.221|:80... 
connected.
HTTP request sent, awaiting response... 404 Not Found
2017-02-03 19:16:27 ERROR 404: Not Found.
===8<------------------------------

The URL you give is unusable, but my attempts to guess the correct URL
also get a 404.  Not that it matters:

> Were you to do that, you would have the freedom to copy, edit and
> share the documentation contained therein. This makes the
> documentation free.

No, it does not.  In your scenario, GNU Radio required *you* sacrifice
essential freedoms (whether you care or not) in order to liberate the
software ad-hoc for others.  The problem is that the distribution
chain still attacks the freedom of /someone/, who then must take a hit
for the team and compromise freedom.

Of course, we can always pay or outsource the freedom loss to a
sell-out (someone who is not aligned with FSF principles) to do the
dirty work.  The problem is freedom is still inherently compromised by
The GNU Radio Foundation, Inc.  Such an effort only serves to further
undermine the free software movement.  It's a disservice to users to
create a mechanism that merely enables a non-free artifact to
masquerade as a "free" one by riding on the weasel-wording that Ian
Kelling and yourself propose as a workaround to free software.

What you're doing is comparable to a middleman buying the eggs of
abused chickens and then relabeling them as "range-free chicken eggs",
to lead consumers to incorrectly believe that their patronage doesn't
contribute to wrong-doing at the top of the supply chain.

Suppose you then rebrand your GNU Radio copy as "Svetlana's Radio",
and then put that in the FSF directory with the official documentation
in the FSF directory pointing to free documention that you distribute.
The problem remains that the GNU Radio record in the FSF Directory
still points to this non-free documentation:

  http://gnuradio.org/

but in a perverse disservice to the users, it is improperly presented
as free documentation.  Svetlana's Radio may be free, but GNU Radio
still is not because the anti-feature flag is currently inconsistent
with the referenced artifacts.  As a remedy, would you be willing to
replace all occurances of http://*gnuradio.org/* in directory.fsf.org
with:

  http://svetlana.nfshost.com/fs/gnuradio-3.7.10.1.tar.gz

so that users are actually directed toward the free version?

> The website you linked is configured badly.

Let's not use oversimplified words like "bad" or "good".
Specifically, gnuradio.org is configured to deny freedom.

> That fact does not interfere with your rights in the case of
> downloading a copy from my website.

Did you add all the official documentation to that tarball (that which
is not included in the original tarball)?  Probably not.

You need to realize that most of the official documentation is not in
the tarball from gnuradio.org.  See my first paragraph here for the
documentation I'm referring to:

  https://lists.gnu.org/archive/html/directory-discuss/2017-01/msg00090.html

If any part of the whole is nonfree, the whole thing is therefore
nonfree.  It's insufficient to redistribute one free small piece of
the whole, and then call the whole thing free.  It's a lie to do so.

I also wouldn't suggest tampering with the original tarball, because
that may be signed by the authors.  Of course a security (trust)
problem emerges if you don't also freely distribute the signature
along with the tarball.  The signature would become unverifiable if
you were to add all the official documentation to that tarball.
Although freedom to be secure is not recognized by the FSF, you still
need (one way or another) to redistribute all the official
documentation in order for your weasel-wording to be defensible in the
slightest.  And that's notwithstanding my other points.


--
Please note this was sent anonymously, so my address will be unusable.
List archives will be monitored.