Re: nongnu Elpa package license requirement: Should it be the other way around?

[Hong Xu]

On 2025-01-13 Mon 01:29 GMT-08, Arsen Arsenović <arsen@aarsen.me> wrote:

> Hong Xu <hong@topbug.net> writes:
>
>> On Sun 2025/01/12 18:18:16-0800 (PST), Richard Stallman wrote:
>>>    > Let's assume a package calls functions from Emacs and depends on Emacs
>>>    > heavily, which is mostly like the case. Should it be required to be
>>>    > licensed under the restriction of being a derivative work of Emacs?
>>> Yes, because they are meant for use combined into one larger program.
>>>    > Practically, this means GNU GPL version 3-(only/or-later) or GNU AGPL
>>>    > version 3-(only/or-later).
>>> Not so.  Many lax, weak licenses are also compatible with those GNU
>>> licenses, and fit the stated requirement.
>>> 
>>
>> Perhaps there's a bit misunderstanding here. Are the packages in
>> non-GNU Elpa considered part of GNU Emacs? If not, how could they be
>> distributed under a permissive license, given that they are linked to
>> and heavily depend on GNU Emacs?
>
> Because the GPL does not require you to release code under the GPL, just
> that the whole combined work can be distributed under the GPL.
>
> This is what is meant by GPL compatibility: if a license allows a piece
> of software licensed using that license to be distributed under the GPL,
> it is GPL-compatible.  No reason to forbid that in NonGNU ELPA.
>
> For instance, an X11-licensed bit of Elisp, hence, is perfectly fine
> to distribute as a combined work with Emacs.  Such X11 licensed code
> could be independently taken and modified under the terms of the X11
> license, and it can also be distributed under terms of the PL as part of
> the 'one larger program'.

The whole reasoning depends on the premise that NonGNU ELPA is seen as
distributed together as combined work with GNU Emacs, even though it is
distributed via network later.

Am I understanding this correctly?

-- 
Hong