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

Re: Swiss Ephemeris Public License



Branden Robinson <branden@debian.org>:

(Big long quote because a few days have passed:)

> On Sat, Oct 11, 2003 at 11:05:56AM +0100, Edmund GRIMLEY EVANS wrote:
> > Branden Robinson <branden@debian.org>:
> > > I personally consider that non-DFSG-free, under the theory that in
> > > general, "your modifications" have pecuniary value, and you are
> > > compelled to license your valuable modifications to the copyright holder
> > > under terms other than those under which you are licensing them to the
> > > community.
> > 
> > What stops you from licensing your valuable modifications under a
> > BSD-like licence so that everyone has them under the same terms?
> 
> There's a difference between "everyone" and "the copyright holder".
> 
> > > Therefore, I see no fundamental difference between this clause and one
> > > which insists that all modifiers pay a license fee to the copyright
> > > holder.  Both cash and copyrightable modifications have pecuniary value.
> > > 
> > > Consequently, in my view, this clause fails the "freely modifiable"
> > > requirement of the FSF's definition of "Free Software".
> > 
> > Would you feel the same way about a licence that said that all
> > modifications must be public-domain or BSD-licensed?
> 
> Public domain?  No, because that doesn't put anyone in a privleged
> position.
> 
> BSD-licensed?  Depends on to whom the license is granted.  If it's a
> public license, then I see no particular problem, though it's a much
> harsher form of copyleft than we're used to.
> 
> I think a case could be made that OpenSSL is in fact already under such
> a license.[1]
> 
> > What about a copyleft licence that grants the DFSG-freedoms but gives
> > additional permissions to Jehova's Witnesses (who happen to come to
> > mind as they turned up at my door as I was typing this)?
> 
> The "additional permissions" test is only applicable if the license
> without the "additional permissions" granted to a certain group or
> individual is already DFSG-free, which is -- I submit -- not the case
> here.
> 
> If the "base license" is not DFSG-free, then the wrinkle that it's
> "DFSG-free" only for a select group of people makes it fail DFSG 5 ("No
> Discrimination Against Persons or Groups").
> 
> Freedom for an elite few is not freedom.
> 
> To boil it down a different way: compelling you to give something away
> to the whole world[2] is DFSG-free; compelling you to give something
> away just to me is not.

I still don't understand this.

Suppose we have:

licence A that forces you to release modifications under a BSD-licence
to the whole world

licence B that forces you to release modifications under a BSD-licence
to the original authors and a GPL-licence to the whole world

licence C that forces you to release modifications under a GPL-licence
to the whole world

Then licence A gives fewer permissions than licence B, and licence B
gives fewer permissions than licence C. If you dual-license something
under A and B that's the same as licensing it under B (because licence
B doesn't stop you from also BSD-licensing your modifications to the
whole world), and if you dual-license something under B and C that's
the same as licensing it under C (because licence C doesn't stop you
from also BSD-licensing your modifications to the original authors).

You seem to be saying that A and C are DFSG-free, but B isn't. So
something released with license A is free, but software dual-licensed
with A and B is non-free. I seem to be seeing or imagining some kind
of paradox here ...

Edmund



Reply to: