Re: [Lynx-dev] Re: GPL

[David Woolley]

> Nonsense.

For a static link, that would fundamentally undermine the GPL.
Static linking is generally understood, by the software industry,
as creating a derivative work.  The contentious area is dynamic
linking.  The FSF view is that dynamic linking should be treated
like static linking, and to achieve a sufficiently arm's length
separation, other components should be self contained programs.
 
> Besides, you can just distribute OpenSSL for Windows and Lynx for
> Windows separately. OpenSSL implements a public interface (and in
> fact GnuTLS speaks some of that as well) and is fairly standard.

This is the sort of brinkmanship that commercial organisations trying
to use GPL code without releasing their code try.  It's probably
never been tested in court, but even if the FSF are wrong about this,
trying it on shows a disrespect for the licence.

If GnuTLS is binary compatible with OpenSSL, then its OK to substitute
OpenSSL, as long as you only redistribute with GnuTLS, but if one
provided a token GPL library that was not realistically complete, it would
generally viewed as an attempt to violate the spirit of the licensing.

> I'd still get the copyright holders (not every single contribution
> is worth enough, you know) permission though.

Yes.  Basically, if the change is trivial and obvious, no copyright
is created, so not every contributor is a copyright holder.