Misunderstanding GPL's terms and conditions as restrictions

[Les Mikesell]

Alexandre Oliva wrote:
> 
>> On Jul 28, 2008, Les Mikesell <lesmikesell at gmail.com> wrote:
>>>> RSAREF didn't stop the program from being created in the first place,
>>>> or from being distributed under the GPL in source form.
> 
>>> Per the FSF, RIPEM was a derived work of gmp and could not be
>>> distributed execept under the GPL.  However that was impossible
>>> because RSAREF was needed and had other terms.
> 
>> I don't see where the FSF said such a thing.  I see the FSF discussing
>> restrictive patent licenses that SSH developers had accepted, and that
>> didn't permit them to distribute the work under the GPL, with as
>> little as plugs for RSAREF to be used.
> 
> Err...  Digging further, I found out that RIPEM included RSAREF, and
> it was RSAREF itself that had been modified in such a way that it
> became a derived work of GPLed work. 

No, RSAREF couldn't have been modified.  It had restricted distribution 
and everyone had to get their own copy. 
http://www.csm.ornl.gov/~dunigan/rsaref.txt
(much later, slightly before the patent ran out, the rsa algorithm was 
release into the public domain)

The linkage to a main program that also linked to gmp is what triggered 
the claim that it was a derived work. It's hard to find stuff this old 
on the internet but I did run across this indication that we are in a 
time warp regarding discussions of what RMS claims are derived works:
http://kerneltrap.org/mailarchive/linux-activists/1993/7/7/36929  (2nd 
hand quote saying you can't run anything but GPL'd code on Linux...).

> So this modified version of
> RSAREF could only be distributed under the GPL, which neither the
> patent license nor the copyright license under which RSAREF were
> provided permitted.

It was the linkage to both libraries that made any usable terms for 
distribution impossible, even though separately both needed libraries 
were available in source.

> So is looks like it is the main program (SSH?) that was the
> distraction.

RIPEM was more of a PGP-alike - ahead of its time: 
http://wiretap.area.com/Gopher/Library/Article/Crypto/ripem.faq.

> Anyhow, if you have pointers to more specific
> information about what the FSF actually wrote (all I found was
> second-hand hearsay), I'd love to know more about it.

I haven't been able to find the after-the-fact summary where the policy 
of considering a work a GPL derivative if it used library functionality 
that was only available under the GPL, but would not be when another 
compatible library became available was explained but I remember seeing 
it.  Here's the usenet thread where Stallman announced the problem and 
an assortment of reactions follow.  In an even stranger time warp, one 
of the comments was mine  - along with some interesting trivia by Alan 
Cox, Dan Bernstein and even Linus.   Ah, nostalgia...
http://groups.google.com/group/gnu.misc.discuss/browse_frm/thread/e117072f9260d78a

-- 
   Les Mikesell
    lesmikesell at gmail.com