[Ffmpeg-cvslog] r5941 - trunk/doc/soc.txt

Michael Niedermayer michaelni
Mon Aug 7 21:24:29 CEST 2006


On Mon, Aug 07, 2006 at 09:29:35AM -0700, Mike Melanson wrote:
> Michael Niedermayer wrote:
> >Hi
> >
> >On Mon, Aug 07, 2006 at 11:37:06AM +0200, Luca Barbato wrote:
> >[...]
> >>>+ Screen the students better regarding their commitments during the
> >>>summer and skills.
> >
> >fully agree, maybe a few questions to check their knowledge of C would
> >be nice before accepting them, that should be done realtime (IRC) to make
> >cheating a little harder
> >some questions about things like !!x, big/little endian, various pointers 
> >to
> >pointers, int32_t vs int, relative speed of * + & % /, 
> >difference of ++i vs. i++, differnt precission of float vs. double, 
> >security
> >issues about checkung input, malloc(x*y) integer overflows, ...
> >ability to work with diff/patch/svn, ...
> >
> >
> >>>+ Provide a more detailed and well-thought-through project list.
> >
> >yes, the whole was started too late before the deadline ...
> >id also suggest that any ideas should be svn commited to our todo list
> >immedeatly so we have a few ideas for the next soc
> >
> >
> >
> >>+ Have weekly reports with a summary of the task done, due next week, in
> >>late
> >
> >IMHO the svn(log) should already be enough
> All good points. Recall, however, that the main reason I rushed to get 
> into this program was to get someone who would do the tedious work of a 
> WMV3 decoder. I think that part has been rather successful.

certainly and maybe that wont be the only part which will be successfull
but i somehow have the feeling that the other projects might take longer
then the timespan for SOC to reach the point of being ok for ffmpeg svn,
iam not sure though due to the changes being generally hard to review
thanks to the cosmetics in them ...
iam also not certain how complete the projects are? maybe someone can
comment here, the ac3 commit messages sounded like its close to being
working, but then it has to be somehow supperior to liba52 and that will
be hard as liba52 isnt crap ...
aac has a big advantage here as libfaad must be dropped due to its license
anyway, so no real need to be better, though being better is of course nice
about amr, well dunno my amr knowledge is loosy ...  but beating the 
reference software should be easy
and then theres vorbis-enc hmm, actually iam curious _where_ it is? theres
nothing in svn ...

Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

In the past you could go to a library and read, borrow or copy any book
Today you'd get arrested for mere telling someone where the library is

More information about the ffmpeg-cvslog mailing list