[FFmpeg-devel] M95 playback system implementation and GSoC questions

[Mike Melanson]

Kevin DuBois wrote:
> 1) M95 is an extension of C93, (it has 5 new compression modes, in addition
> to c93's) so I am wondering whether I should try to cram everything for m95
> decoding into libavcodec/c93.c or to make a new file, like libavcodec/m95.c
> . The second option is easier insofar as I understand the API, but the first
> one will avoid code duplication. If anyone has either has any useful
> suggestions on how to fit everything into one file or endorses the addition
> of libavcodec/m95.c, feedback would be appreciated!

I think you could safely extend the C93 decoder in the same file.
Michael frowns upon code duplication in any form.

> 2) Is there a hard deadline for the completion of the work? I've been aiming
> to have it done by March 31 (the last day for student applications) or
> sooner...

Google stops accepting applications on March 31. But the "interim
period" on the official schedule lasts until April 11. That's when we
mentors need to have final rankings for the applications. Ergo, you have
until around April 10 to get the qual task completed.

That said, keep in mind that "completed" means that the code is in SVN.
Before the code gets to SVN, it needs to be thrashed around here on the
list for several iterations, guaranteed.

> 3) As for a main project, I was considering making streaming support usable
> for the snow codec via rtp or helping to standardize the codec to make it
> 'stable'. Are either of these projects appropriate lengths for a gsoc
> project? I know Marco Gerards was talking about snow improvements, perhaps a
> small group of GsoC students can work on making snow a less experimental
> codec?

Will defer to Michael on this...

-- 
	-Mike Melanson