Re: [Lame-dev] LAME license
Reinhard Tartler a écrit :
b) Additionally, there is a README file included, which contains this
text:
,----[ taken from LAME's README file:
| This code is distributed under the GNU LESSER PUBLIC LICENSE
| (LGPL, see www.gnu.org) with the following modification:
|
| 1. If you determine that distribution of LAME requires a patent license,
| and you obtain a patent license, you may distribute LAME even though
| redistribution of LAME may also require a patent license.
|
| 2. You agree not to enforce any patent claims for any aspect of
| MPEG audio compression, or any other techniques contained in
| the LAME source code.
`----
Disclaimer:
What I am writing within this mail is only my sole opinion.
I am not qualified to give any legal advice, and my own opinion should
in no way be considered to be the one of the LAME project as a whole
entity.
> Now the member of the ftpteam is concerned that this might be an
> additional restriction that is not allowed by the (L)GPLv2.
LGPLv2 doesn't allow any licensee to set additionnal restrictions. However,
the original licensor (ie the copyright holder) is always free to set any
further restriction.
Thus, to me this license is valid, even if this is not the usual LGPLv2
license. (but there is a twist, see below)
Since there was no really helpful response on that inquiry, I've tried
asking at FFmpeg upstream, which provide software that can optionally
use LAME to encode to the MP3 format. On that mailing list, Reimar
Doeffinger points out that because this additional restriction is not
visible in the respective source files but hidden in the README, this
"almost certainly wouldn't hold up in any court":
http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/115654
Do you agree with this claim?
Without ressorting to any court, I share the opinion that every file
featuring the usual LPGLv2 header is only convered by the LPGLv2,
without any consideration of the two problematic notes located within
the readme file.
To me, those should only apply to "headerless" source files. It is even
likely that several of us haven't even noticed/considered those two
notes when commiting new files, faithfully willing to commit new files
only under LGPLv2. (at least that is the case for me).
As a last resort, I think we can still redistribute lame under the terms
of the LGPLv3, which seems the be permitted by the LAME license. The
GPLv3 contains terms that are very similar to the modifications quoted
above in §11. This should address the "additional restrictions" concern.
Do you agree with this theory?
I think that you can not redistribute LAME only under LGPLv3. According
to LGPLv2 you can choose to apply LGPLv3 to LAME, but you should not be
able to remove its LGPLv2-ability.
Now, let's have a look back at those two notes:
They have been written 10 years ago by Mark Taylor. Considering that
every source file from LAME is including the usual, unmodified, LGPLv2
header, I think that unmodified LGPLv2 applies to every source file.
Those notes should then only apply to "headerless" files (ie mainly
documentation).
It would then be reasonable to ask Mark if he agrees to drop those two
notes, and I think that it is likely that the LAME project would reach
an agreement about those.
BTW, while the note #2 is an "activist" one, note #1 is a "pragmatic
one". LGPLv2 theoritically prevents any patent holder or patent to
distribute LAME. That is a huge problem for some companies, and goes far
beyond only LAME.
Note #1 looks to be a way to circumvent this without contradicting the
license.
Please note that I have absolutely no will to go against any patent
holder or patent licensee which would want to use/distribute LAME. As
long as they respect the other LGPL provisions (source code
availability, contributions back, disclosing the use of LAME), they have
my bless.
--
Gabriel Bouvigne
www.mp3-tech.org
personal page: http://gabriel.mp3-tech.org
Reply to: