Re: Licensing of libposix

[Bruno Haible]

Eric Blake wrote:
> In fact, given the earlier question about libposix (should it
> be LGPLv2+ or LGPLv3+), we may be repeating our line of questioning.

Paolo Bonzini wrote:
> In fact, shouldn't we aim at LGPLv2+ for libposix, since that's the 
> glibc license?

Because of the known advantages of LGPLv3 over LGPLv2 (internationally
safe wording, handling of patent threats, rejection of vendor-controlled
hardware like iPhone or iPad, etc.) I think we should use LGPLv3+, not
LGPLv2+, whenever possible.

Our current uses of LGPLv2+ come out of the desire to be helpful to
packages that were using gnulib before 2007, and to packages to which
Eric or Jim contribute. These packages are using gnulib via gnulib-tool,
that is, by inclusion of source.

When packages will use libposix, they will not do so by inclusion of
source code, but by distinct distribution and runtime linking. Therefore
since we already agreed that libposix' license should be in the LGPL camp,
any package can link to libposix.

In the LGPL camp, we have
  - LGPLv3+,
  - LGPLv3+ | GPLv2+,
  - LGPLv2+.
LGPLv3+ has the problem that GPLv2 programs cannot link with it, which
denies GPLv2 programs a privilege that is granted to proprietary programs.
Therefore the FSF recommends to use LGPLv3+ | GPLv2+ in this situation [1].

Whereas I don't see any benefits of LGPLv2+ compared to LGPLv3+ | GPLv2+.

So, my vote for the license of libposix would be LGPLv3+ | GPLv2+.

Bruno

[1] http://www.gnu.org/prep/maintain/maintain.html#Licensing-of-GNU-Packages
-- 
In memoriam The victims of the Skelani massacre 
<http://books.google.de/books?id=nltdtAo38K0C&pg=PA41&lpg=PA41&source=bl&ots=DX2VTN1nga&sig=d_BXUZM0I81hZtfrSBA3392p-KY&hl=de&ei=JHy6TcGTCYmLswbpsrDzBQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CBkQ6AEwAA#v=onepage&q&f=false>