Re: tunefl and other web services

[David Kastrup]

Joseph Rushton Wakeling <address@hidden> writes:

> On 07/07/12 06:10, David Kastrup wrote:
>> This is not a question of reinterpretation or optional.
>
> When I referred to being able to interpret a "GPLv2 or later" work as
> GPLv3, it's not _my_ interpretation -- it's an explicit permission
> granted by the wording of the license grant.
>
>     This program is free software; you can redistribute it and/or
>     modify it under the terms of the GNU General Public License
>     as published by the Free Software Foundation; either version 2
>     of the License, or (at your option) any later version.
>
> That's an explicit grant that says you can take _either_ the terms of
> the GPLv2 _or_ the terms of a later GPL version, and you can pass to
> downstream users either set of permissions for your modified version.

But not something which would require cherry-picking some terms from one
license, and some terms from another.

> Hence, you can link "GPLv2 or later"-licensed code against
> GPLv3-licensed code, so long as the work as a whole is considered to
> be covered by GPLv3.

But you can't link "GPLv2 or later" code with both GPLv2-only and
GPLv3-code, and once you link with either, the choice is fixed.

> I'm not advocating for AGPLv3; it's just that I can't see a GPLv2
> compatibility case for AGPLv3 that doesn't also apply to LilyPond's
> current licensing choice of GPLv3.

AGPLv3 basically is a GPLv3 variant with additional restrictions.  The
boilerplate GPLv3 has an explicit clause allowing to combine GPLv3 and
AGPLv3 parts into a program covered by AGPLv3.  That is more or less a
one-way street.

Putting LilyPond under the AGPL does not buy much: people offering a
network service based on LilyPond need to make the source of their
version of LilyPond available for download.  I don't see any evidence of
modified versions of LilyPond being in use.  Switching to a lesser known
license does not seem useful for me at this point of time.

-- 
David Kastrup