[mythtv] More scheduling scheduler

Max Barry mythtv at maxbarry.com
Sat Apr 15 09:23:50 UTC 2006

David Engel wrote:
> You guys try it both ways, with and without softpad always considered, and
> let me know which way you like better. When you decide, please
> provide a clear rationale. Those of us who sitck around and have to
> support it (Bruce, Michael, myself, etc.) can then regurgitate it when
> the questions come up again in the futrue. 

I haven't tested the code yet (am preparing the box now!), but I know
I'm on Paul's side here. Softpad should absolutely not be canceled by a
hardpad. Here are my hopefully-regurgitatable reasons why:

First, we're calling this a global buffer. If it's global, it should be
applied... well, globally.

Second, hard and soft buffers serve very different purposes. Hard
buffers are for correcting known differences between the guide data and
the actual schedule time. Soft buffers, on the other hand, are insurance
against unpredictable time shifts.

An example: say I have a global 5-minute end-late soft pad--maybe I
don't trust my system clock, or the network's. Say also that although
SURVIVOR is scheduled for 8-9pm each week, I have noticed that the
credits always run a minute longer. That is, the show is actually 61
minutes long.

What I should do is simply extend the length of SURVIVOR by one minute.
I am thus using what I know to correct the guide data. But I am *not*
also saying I no longer wish to insure against clock drift.

If my action doesn't affect global softpad, everything is fine. MythTV
will record from 8pm to 9.06pm if it can honor the soft buffer, and 8pm
to 9.01pm if it can't.

If, on the other hand, this cancels the global softpad, MythTV will only
ever record to 9.01pm. Suddenly I am no longer insured against clock
drift. Even if I'm aware of this, I have no sensible way of
compensating. I could make that hardpad 6 minutes, but that's not the
same, because my "just-in-case" soft pad has now become a
"always-record-even-if-it-means-conflicting-other-shows" hard pad.
That's not what I wanted.

Although they both affect the precise time MythTV will be recording,
soft padding and hard padding have very little to do with each other

Thanks again for the code--and the summary! Looking forward to testing
this. It sounds great.


More information about the mythtv-dev mailing list