[FFmpeg-devel] Areas needing work

[Stefano Sabatini]

On date Tuesday 2015-04-07 14:40:50 +0100, Derek Buitenhuis encoded:
> On 4/6/2015 11:23 PM, compn wrote:
> > i think we tried this a few years ago and found that we do still have
> > ffserver users, they are just quiet until we threaten to rm it. if/when
> > it breaks, someone will either be there to fix it or it will rot.
> 

> That's stupid. Either someone steps up to fix it (read: actually cares),
> or it remove it. Having a broken piece of crap that nobody cares about
> in the repo benefits nobody.

I did some consulting about the use of ffserver in the past, with
variying degrees of success (hint: RTP streaming is usually tricky and
highly depends on network and bandwith constraints). I agree that
ffserver is painfully underkill, but still better than nothing. Also
we have an official maintainer and some substantial work has been done
on it recently.

Use cases: real time streaming of interactive content through HTTP or
RTSP/RTP. Probably these days there are better solutions but I think
ffserver still cover an use case which cannot be fullfilled by the
other FFmpeg tools. At the present days there are probably other valid
alternatives (e.g. Flash/RTMP, webRTC, HTML5/Media Source Extensions,
JavaScript native decoding + websockets, HLS, DASH) which cover
similar areas, but are not completely equivalent. For example HLS is
known to introduce some significative delay.
-- 
FFmpeg = Fundamentalist & Fiendish Multimedia Pitiful Elfic Game