dragonfly license

Rahul Siddharthan rsidd at online.fr
Sun Oct 3 22:27:52 PDT 2004

Matthew Dillon wrote:
>    It's not silly, but it does point out a serious flaw with GNU.  In
>    recent years GNU has tried to create a 'floating' copyright.  That is,
>    one where the code simply references some ephermal standard gnu copyright
>    residing somewhere outside the file being copyrighten.
>    This is very dangerous, because there is no court precedent for allowing
>    a published work's copyright to change after the fact and no way to
>    determine, short of recording an exact date and version (and hoping that
>    the version is properly updated on the site), which copyright the source
>    actually refers to.

Actually, the GNU GPL is versioned: the present version is 2, and the
text of this version is sufficiently widely distributed that there's
no way the FSF could retract or modify it or deny its existence even
if they wanted to, so I don't think this is an issue.  The copyright,
of course, should be in the file itself, eg "Copyright 1994 John Doe;
redistribution permitted under version 2 of the GNU General Public
License", with instructions on where to find the text of the GPL.

The FSF recommends to software authors (and many obey) that they allow
redistribution under the GPL version 2 "or, at your option, any later
version".  But this is not necessary, and Linus Torvalds for example
rejects it: the linux kernel may be distributed only under the GPL
version 2, not under any later version.


More information about the Kernel mailing list