gpgme license

marius aamodt eriksen marius at
Fri Aug 16 07:18:01 CEST 2002

* Marcus Brinkmann <Marcus.Brinkmann at> [020815 23:23]:

> On Wed, Jul 24, 2002 at 12:57:22PM -0400, Niels Provos wrote:
> > He would like to link his BSD-licensed code against gpgme.  As gpgme
> > is GPL, that would make his code automatically GPL, too.
> As was repeatedly said in this thread, this is not true.  The GPL applies to
> the combined work, he can still choose to dual license his code under
> whatever other licenses he wants to provide it under.

but my code would be unusable if it isn't combined.  so effectively it
is GPLed.

furthermore, i have spoken to many people about this since our last
encounter - people that are heavily involved in open source software,
and they all agree with me.

> > Sorry, your indoctrination by Stallman seems not yet complete.  Marius
> > is well known for developing and releasing free software.  Calling him
> > a hoarder seems very undiplomatic.
> Although I can see how you can read this into what I wrote, my intention was
> not about Marius (please read again the sentence in parenthesis).  The
> problem with the LPGL and other less restrictive licenses is that they help
> hoarders, whoever they are.

i assume by hoarders, you mean people who use your software, what's
bad about people using your software?

> Your criticism doesn't reach its target, for a simple reason.  There is no
> conflict between the modified BSD license and the GPL.  If all you want is
> to combine two such works, you are free to do so, and we do it in many
> projects with a huge success.  

again, as i stated before, this seems to be untrue; if it WAS true,
then why was ximian forced to use another solution when they needed
PGP integration in evolution?  they were not interesting in modifying
gpgme, only in using it for a solution which could have commercial
extensions.  to me, this seems to be the exact same situation; no
conflict by your argument, but ximian somehow, who i am sure have a
very careful legal department, decided they could not use gpgme, go

also, FYI - since our last encounter, i have implemented my own PGP
API, named libpgpwrap; it is modular, flexible and simple.  it
supports a number of pgp implementations as its backend (currently
PGP5, GnuPG in the works, PGP2 next).  honestly i believe the
interface i have come up with is superior to gpgme's; it is intuitive
and significantly less complex than gpgme's.  using pgpwrap i can
implement a full PGP client (sign, encrypt, verify) in about 80 lines
of code.  gpgme seems to have overdone a few things, and libpgpwrap
simplifies without compromising functionality of flexibility.  and,
after pgpwrap matures, it is possible i will use tom zerucha's openpgp
implementation (which i can distribute with libpgpwrap) as a backend;
making pgpwrap completely independent of any other implementation.
libpgpwrap, of course, is under a BSD style license.

perhaps this is a demonstration of how truly free software will evolve


> marius at >

More information about the Gnupg-devel mailing list