[mythtv-users] possible p2p approach for mythtv information?

[Brad Templeton]

On Wed, Jul 05, 2006 at 07:26:51PM -0500, Kevin Kuphal wrote:
> Simon Kenyon wrote:
> > On Wednesday 05 July 2006 21:14, Brad Templeton wrote:
> >   
> >> I don't see this as a P2P app.  First of all, I suspect the vast majority
> >> of Myth boxes are behind NAT and would not readily participate in a
> >> P2P app.    Nor would you necessarily want to gather info by allowing
> >> many thousands of machines to connect to you.
> >>     
> >
> > generating cutlists using some sort of p2p would seem to me to be the correct 
> > way to do it. after all, who is going to provide a central infrastructure? 
> > and for what gain?
> >
> >   
> I have a hard time even seeing the value of sharing this data.  Cutlists 
> are quite dependent on pre/post-roll, local commerical segments and many 
> other factors making it a losing battle to even collect valid 
> information and then, for what?  My commercial detection generates a 
> near perfect cut/skip-list for me within an hour of the completion of 
> the recording and I rarely have anything in my watch list for more than 
> days before it is watched and deleted.

It is imagined that one day there will be a battle between networks/advertisers
and commercial eliminators.    Shared cutlists create a system that will
always win, you can't defeat thousands of human intelligences.

It is a challenging project, though it's vastly easier on programs that
have closed captioning, because then you can sync up bits of the show
with the text that is known to be in the show or known to be in commercials.

(In fact, when commercials are captioned, it allows a very clever commercial
skip algorithm because you can detect their text and compare it to a database
of known commercials.  If they are not captioned, it makes spotting them
even easier when the program is.)

If neither have CC, it should still be possible to generate signatures for
programming based on metrics such as scene duration -- the sequence of
gaps between total cuts/scene changes.

Note that it can often be simplest just to have a large user base so there
are many people watching the exact same broadcast, which is much easier with
satellite channels of course.

However, this is still a fancy problem.  It's not likely to be the
highest priority for myth coders.    Sharing recommendations on programs
that suck or are good is another matter, much more doable.