[MPlayer-dev-eng] [RFC] Some modest proposals

[Dan Oscarsson]

On 2010-06-12 at 14:34 +0200 Luca Barbato wrote:
> 
> > not that i truely care one way or the other. i just think some reasons
> > are being blown up. i mean if you want to switch to git because you
> > like it more, thats a better reason to me.
> 
> As I said I was wondering why MPlayer is less fun to hack and those are
> my reasons.

I have now been hacking on mplayer for one and a half year. Slowly I
begin to understand parts of it. Having written a lot of new code to
mplayer.c - it have been times when I really have wanted to replace all
of it with new code.
To make hacking mplayer more fun I would like to see mplayer as taking
the leading role in modern media players. Today mplayer feels somewhat
ancient and I get the feeling that it is loosing in popularity. VLC is
going strong.
The code feels full of ancient parts that very few can need today.
While mplayer have been fairly easy to modify to make it play movies
very well on LCD-TVs, I would have liked very much to work on a mplayer
2.0 where much of current core is scrapped. I would expect threads and
parallell demuxing/decoding/displaying to efficiently use the modern
CPUs. Maybe core running as a daemon, like XMMS2, with a standard
communication library (instead of slave mode) to create users interfaces
with. And a modern user interface.

Many people think that now and then, you have to scrap most of the old
code and start again. mplayer shows it age - a rewrite of the core could
give a new start.

That I expect would make hacking mplayer more fun and draw new
developers.

   Dan