[mythtv] RFC on proposed scheduler patch re: hdtv dups

[f-myth-users at media.mit.edu]

    > Date: Sat, 28 Jul 2007 11:42:33 -0700
    > From: Bruce Markey <bjm at lvcm.com>

    > I'd think a better approach would be to have a per rule option
    > to double record it's episodes with the provision that the two
    > recordings must be from a different sourceid if the episode is
    > on more than one source else a different chanid if the episode
    > is on more than one channel (frequency) or at least a different
    > input if it is only available from one chanid.

That's a really powerful idea.

It would certainly help many of the issues I often confront, e.g.,
something's repeated 15 times on PBS, but available from 3 PBS
sources, and sometimes one of the sources is fuzzy because their
studio-to-transmitter link has problems in the rain---grab two from
dissimilar sources automatically and sort it out later.  (Otherwise,
the options are "remember to double-record" which is what the computer
should be doing for me, ro "record every showing" which records 10x as
many copies as I need and/or causes too many scheduling conflicts.)

Or grab two copies of the Daily Show from its 4-5 repeats/day, just in
case one of them got an emergency-alert-test in the middle, without
having to go in and specify -which- two (which tends to get in the way
of allowing the scheduler to juggle things around).

Even -with- don't-lock-on-dumping-ivtv-buffers-when-scheduling patch
of early March, I still sometimes get (much less noticeable) glitches
when the scheduler fires; having a couple recordings of something lets
me choose the one that didn't have the glitches.  (Of course, the real
solution here is to fix whatever's still causing lossage, but...)

I'm sure there are many other examples where this could be handy.

In general, the ability to "record any n of these" instead of "record
just 1, or record -all-", would be handy.  Going for input diversity
first is of course the right idea, and falling back to "just record
multiple copies even if you have to do it from the same input and
chanid" works for cases where all your inputs are "the same" but you
still want to record more than one copy in case something goes wrong.
(Of course, unless you have ample tuners, you might want to say "but
make new recordings always win over redundant recordings" or something.)