[MPlayer-G2-dev] the awakening, license changes and so on...

[D Richard Felker III]

On Wed, Aug 04, 2004 at 03:54:04PM +0200, Diego Biurrun wrote:
> Alexander Strasser writes:
> > On Wed, Aug 04, 2004 at 09:51:46AM +0900, Attila Kinali wrote:
> > > > - getting g2 api accepted as 'industial standard for linux/unix video/media'
> > > >   something like oms, and later gstreamer and openquicktime wanted to
> > > >   reach, with no much success.
> > > >   we have to make it available for commercial users, otherwise they wont
> > > >   spend their time developing/porting their codecs, demuxers etc for it.
> > > > - getting some money/hw/sponsorship to developers who need it to be able
> > > >   to work on the code fulltime or at least more time.
> > > 
> > > Why do you want to get commercial users onto MPlayer ?
> > > Just for the sponsoring ? 
> > Don't forget it be should the movie framework for ALL. Think about computer
> > games and lot's of other out of movie playback applications ( as the
> > main focus is on something else ), that can't publish their code, using
> > MPlayer G2 for movie playback and what not.
> 
> Games could use GPL MPlayer G2 for movie playback without having to
> publish their code.

First of all, hi guys, I'm back.

I know this is old, but I'd like to stick a few comments in here...
Something like MPlayer (g1 or g2) is completely useless to a game
wanting to play movies for cut scenes or whatever. The whole purpose
of a program like MPlayer is its versatility. The ability to play any
random-ass broken movie file, filter to correct for problems in the
source or to adapt it to the format you want to view it in, seek back
and forth, etc... A game will be showing movies of exactly one file
format & codec, will be playing them linearly (no seeking), and will
not need any filtering or whatnot. A very very basic player core is
much more appropriate for this than a 10-20 meg MPlayer.

Now, the same is true of lots of other proposed uses: cable tv video
on demand, kiosks, huge screen displays at events, etc. Whenever you
have homogeneous data to play, homogeneous display requirements, etc.,
you can get by with a really simple player tailored to your purpose.
No MPlayer needed.

While I was away I learned about Arpi rejoining G2 and his LGPL plans,
and I decided way back then not to flame. I'm tired of seeing everyone
mad at each other. I've also realized that Arpi's vision of G2 is a
lot different from what I want(ed) to work on. As long as no one
commandeers GPL code into G2 without the authors' permission, I have
no objection to the G2 authors (Arpi, etc.) making the core LGPL. I
think it would be more strategic to stick with GPL, but it's not my
business. I hope they'll be open to having things outside the core be
GPL and having the "default" player app be GPL, but this isn't
important right now.

I'll continue to work on some code with MPlayer (G1 and G2) under GPL,
mainly stuff I've already been doing like mencoder and video filters.
Maybe introducing a new audio filter design too, and maybe (_maybe_)
even giving a little of this code to G2 under LGPL if it needs to be
to get in G2. In the immediate future I won't have a lot of time for
this anyway though.

Right now my priority with respect to MPlayer/ffmpeg will be
finalizing the NUT spec and getting NUT deployed. Along with snow and
sonic, I think we have the chance to take a much bigger place in
(crossplatform) multimedia. In the future, I still want to work on
something like what I proposed before, a new framework for audio and
video that's much more radical than what Arpi has in mind for G2 (i.e.
what he dubbs "G3"). If I do this, it will be under GPL, and will fill
a comparable role to something like gstreamer or DirectShow except it
will be done right.

Well that's all for now. Sorry to disappoint with no flames, and glad
to have everyone back on the team.

Rich