[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

cypherpunk license: PLEASE STEAL ME (Re: GPL & commercial software, the critical distinction) (fwd)




Forwarded message:

> Date: Tue, 29 Sep 1998 01:53:23 +0100
> From: Adam Back <[email protected]>
> Subject: cypherpunk license: PLEASE STEAL ME (Re: GPL & commercial software, the critical distinction)

> OK, but it would seem to me that to interpret GPL in this way is in
> violation of the spirit (and probably legal intent if interpreted by a
> lawyer).

Actualy this situation is specificaly addressed in the LGPL in regards
header files and how they are the gray area in all this and probably
indefensible in a court (it says this in the LGPL go look...www.gnu.org).
As programmers this is what we want after all, the API to be described so we
can replace those proprietary libraries and programs if we desire.

Actualy you missed the obvious way to distribute commercial code and the
(L)GPL will protect your propietary work....

In both licenses it is very specific in noting that a L/GPL'ed work that is
subsumed in another causes that work to become L/GPL'ed by default. It
further notes that there is NO implication that the use of commercial
code/libraries in a L/GPL'ed work can be construed to L/GPL that proprietary
code.

How I do jobs for my customers is that I develop my programs and libraries I
want to protect via copyright and then distribute them with a set of scripts
and makefiles that during install build the end product. So the end user can
enjoy the privileges of the L/GPL'ed code (ie can fix bugs and relink the
libraries to their hearts content) and I protect my work. The catch is to
make sure the makefiles and scripts are L/GPL'ed. You could even include gcc
or egc on the distribution medium and use it to compile the resultant
without worrying about your commercial code being subsumed. It is quite
commen today for commercial houses to use gcc/egc to develop and distribute
fully commercial code, the catch is that none of the end resultant code can
come from L/GPL'ed sources.

This harks back to my comment earlier today about the GNU/cash system being
a perfect medium for the implimentation of crypto/e$ protocols. The catch is
that you have to be willing to give up your claims to the stubs that must be
placed in the L/GPL'ed product. If you can get away with using named pipes
as buffers and T's then it becomes reasonably trivial to create a situation
where one can inject non-L/GPL'ed code into the stream without the
functional code so injected falling under the L/GPL.


    ____________________________________________________________________

                            The seeker is a finder.

                                     Ancient Persian Proverb

       The Armadillo Group       ,::////;::-.          James Choate
       Austin, Tx               /:'///// ``::>/|/      [email protected]
       www.ssz.com            .',  ||||    `/( e\      512-451-7087
                           -====~~mm-'`-```-mm --'-
    --------------------------------------------------------------------