Read and respond to this message at:
By: adfhau

Yes, things are going to vary from protocol to protocol..

Though one would assume there are standard components to dealing with a webcam
in sending and receiving..

Ie. A - Sending

1. Fetch stream from camera - assuming V4L or V4L2? What other video protocols
are there besides devices such as DVB-x receivers which stream MPEG anyway or
IEEE1394 devices?
2. Pass stream from camera to codec + display routines (w/display 
like Xv and friends - perhaps xine and other programs have display libraries
that would be helpful here?)
3. Pass codec output to IM protocol video protocol handler
4. Pass IM protocol video protocol handler data to core IM protocol handler

... conversely ... B - Receiving

1. Receive stream from core IM protocol handler and pass to video protocol
2. Pass stream from video protocol handler to codec
3. Pass from codec to display routines

Obviously there's going to be a lot of variation here - but has the framework
for say steps A1-2 and B1-2 been done? I would think that things like ffmpeg
and friends may have done some of the codec work.. in line with my previous
comments, the telling thing here will be to see which one gets video first..
Skype for *nix or Gaim for Linux..

Now.. I'd contribute if I could - however my C and C++ coding is shite..
malloc/dealloc/realloc always did my head in, and my university seemed to focus
on java..

I wonder what's happening with gaim development - is it that the major protocols
are changing their authentication and checking mechanisms so often that all
the work is going into reverse engineering them? Are there patent issues on
the codecs used to decode/encode video streams? I guess in some ways it's also
telling that file transfer doesn't work reliably on some of the larger

Perhaps someone needs to, if they haven't already, design the kind of framework
above if it hasn't been done, and say "These are your APIs" and then those who
were capable could just focus on taking stream A and sending it to socket B..
I dunno.. maybe I'm waaay too simplistic about things. I just get the impression
that often in open source stuff that people insist on reinventing large swathes
of code because something could be done better at the expense of zero progress
because the same ground is being recovered multiple times.

You are receiving this email because you elected to monitor this forum.
To stop monitoring this forum, login to and visit:

Take Surveys. Earn Cash. Influence the Future of IT
Join's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
Gaim-forums mailing list

Reply via email to