Re: [Chicken-hackers] Re: Backdoor GPL in message-digest

[Magnus Achim Deininger]

On Tue, 24 Aug 2010 20:16:58 +0200, John Cowan <address@hidden> wrote:

> On Tue, Aug 24, 2010 at 1:29 PM, Peter Bex <address@hidden> wrote:
>
> > That's not what I said.  I said you can't link a LGPL library to a GPL
> > library.  You said you *can* link a GPL library to a LGPL library, which
> > is of course correct (but the other way around).
>
> Ah, I see: we are using the term "link to" in different senses.

Incidentally, this is precisely the major problem with both GPL and LGPL:  
defining what actually counts as linking and what does not. I remember a  
discussion on some other list quite a while back that would recommend not  
using either licence for Scheme or LISP code because with a lot of LISPs  
the distinction between the compilation, interpretation and linking phases  
did not apply in the same sense as with the languages that had been in  
mind when designing the GPL (mostly C and derivatives, it would seem). The  
thread ended in strongly advising people to consider BSD-2 or X/MIT  
licences instead, since they don't rely on those particular compiler  
semantics in any way.
A similar problem also applies to languages like Pascal, where the  
interface description for individual modules is contained within the  
unit's source file. It's fairly typical for open source Pascal code to be  
redistributed and included in projects by just copying the needed source  
file directly, as opposed to making shared libraries and distributing  
headers (although that is possible in some dialects, it's mostly too much  
of a hassle).
Basically in languages like that, the LGPL becomes identical to the GPL  
since there is no distinct linking stage that could be performed  
separately.

> ...

Cheers,
Magnus