[mythtv-users] Integrating non-traditional program sources into MythTV (won't tell you what it 'was:')

David Shay david at shay.net
Wed Jul 23 05:14:31 UTC 2008


On Tue, Jul 22, 2008 at 5:05 PM, Jay R. Ashworth <jra at baylink.com> wrote:
> On Tue, Jul 22, 2008 at 02:55:23PM -0500, Kevin Kuphal wrote:
>>    It is an interesting conundrum. It really can't be integrated
>>    into searches per se because it isn't guide data.
>
> Expand on the assumptions you make within "it isn't guide data", would
> you?
>
>>                                                         It doesn't fit
>>    in the traditional role of the program guide because the programs
>>    are instantly available without notice. How did you envision this
>>    appearing? Would you expect to see feed "recording rules" along
>>    side normal rules in the Recording Priorities list?
>
> Well, I would just treat them as channels, each with it's own
> (pseudo-)tuner, and the "future data" is just the available programs.
>
> Remember: some RSS feeds only show the latest items, but items
> announced earlier can still be retrieved.
>
> So given that, I don't think it's unreasonable to merely simulate a
> program guide, dropping each new item onto the grid for it's 'channel'
> as of when you first see it, and droping multiple items in sequence if
> more than one thing appears simultaneously.
>

To me, this analogy breaks down in two key ways.  From a UI
standpoint, made-up times just would seem kind of strange.  Do you
just list all available content with invented times starting from
right now, without repeats, then just stop the grid when you run out
of content?  Do you use the "air" date from the content (but that
would always be in the past...). Seems kind of odd/incongrous.

Then, from a code standpoint, this is very much
totally-unscheduler-like.  Having spent a few days in the scheduler
code working on a patch for something else, I can confidently say that
this kind of thing would be a total rewrite of the pretty much the
entire scheduler.  It is completely reliant on times and dates as a
key component of its logic, and making ones up just to effectively do
some regexp searching and intelligent downloading/DB inserting seems
like overkill.

I can see a different type of search screens being setup, with some
similar constructs around searching by title, and some similar things
like duplicate checking, but it's still enough different from a
scheduled broadcast model to require a different backend data model
and GUI as well, in my opinion.


More information about the mythtv-users mailing list