[FFmpeg-devel] Support master branch of OpenJPEG and Grok J2K codecs

[Aaron Boxer]

On Mon, Apr 4, 2016 at 6:54 PM, Timothy Gu <timothygu99 at gmail.com> wrote:

> Hi,
>
> On Mon, Apr 04, 2016 at 03:54:11PM -0400, Aaron Boxer wrote:
> > >
> > > >
> > > > The following changes were made:
> > > >
> > > > 1. Removed bpp (redundant as this information is already stored in
> > > precision)
> > >
> > > Does compilation still work without this change?
> > >
> >
> > Yes.
>
> Then the change is unrelated, and either needs to be in a separate patch or
> not commited at all.
>
> > >
> > > Does the compilation of OpenJPEG break _now_ without the patch? If so,
> > > submit
> > > a bug report, if not, then this change merits further discussion.
> > >
> >
> > Compilation against OpenJPEG master is broken, but everything works with
> > the latest release
> > of OpenJPEG (2.1)
>
> Would be great if you post the compilation log and file a ticket.
>
> > >
> > > If there isn't a way to detect Grok from OpenJPEG, there should be one.
> > >
> > > If it is not clear to you why we are so against AGPL, it is because it
> > > incurs
> > > additional restrictions on the work that we don't consider to be in the
> > > spirit
> > > of free software, regardless of what FSF says. But I think you already
> know
> > > about that.
> > >
> >
> > Why do you consider it to not be in the spirit of free software?
> > So far, nobody has given me a convincing argument against the use of  the
> > AGPL.
>
> I believe most members of the FFmpeg community consider free software the
> same
> way Linus Torvalds (among others) considers it: share the
> sources/modifications if you are distributing them, and do whatever you
> want
> with it if you are not.
>

Yes. Well, AGPL extends the definition of distribution to include use over
the network.


>
> We have no problem whatsoever with the entities using FFmpeg on their
> server
> without distributing the binaries or releasing their sources, or
> tivoization
> (which is a related, but different issue). Sure, it's better if they submit
> the patches, and we think that they are missing out by not submitting them,
> but we don't have any problems with it.
>

Fair enough.

>
> We do have a problem with the people that are preventing others from using
> the
> software under the aforementioned legitimate circumstances. For existing
> non-free software, we really don't have a choice, and it is clear to the
> user
> that such a copy of FFmpeg is non-free. But licenses like AGPL which claim
> it
> is "free" but fail our criteria of freedom are deemed to be "evil." In
> fact,
> AGPL is much more restrictive than most non-free licenses we deal with,
> since
> it concerns use in addition to distribution (see also Reimar's comments).
>

I see. So, licenses that fail your criteria of freedom are considered
"evil".
That is a pretty black and white view of the software world.
Also, I would point out that proprietary licenses most definitely restrict
use in addition to distribution.


>
> Moving forward, thanks to the explicit GPL compatibility shoehorned into
> Chapter 13 of AGPLv3, there isn't anything that makes us license our _own
> work_ as AGPL simply because FFmpeg binary is linked to an AGPL work (in
> fact,
> we cannot relicense our own code from LGPL to AGPLv3 like we can from LGPL
> to
> GPL). But using such a resultant mixture of licenses isn't something we
> want
> our users, personal or enterprise, to deal with.
>
>
This is a bit like parents thinking if they don't tell their kids about
sex, then they won't have sex.
We all know how that turns out.



> And that is why we are making more fuss about AGPL than most non-free
> libraries.
>
> Well, I don't want to flog a dead horse, so I guess we can all end the
discussion here:

Interesting points were raised, and  I learned some new things about
licensing.

Cheers,
Aaron





> Timothy
>
> P.S. I am not a lawyer.
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>