[mythtv-users] zap2it guide data wrong in myth, correct in zap2it site?
Michael T. Dean
mtdean at thirdcontact.com
Fri Oct 20 06:11:41 UTC 2006
On 10/20/06 00:32, Marco Nelissen wrote:
>>>The bandwidth isn't the constraining factor for TMS. It's server
>>>(computational) resources. When every single DataDirect user has a
>>>unique lineup, creating daily hashes requires /more/ resources than just
>>>downloading again. However, there's no nead to download again because
>>>Myth does it right.
>>>
>>>
>>Makes sense,
>>
>>
>Not really, I think. First off, not every user has a unique channel
>lineup of course. There will be plenty of people who have the exact
>same lineup. Furthermore, just because I disabled some channels in
>my lineup doesn't mean that they *have* to treat it as a different
>lineup requiring a different hash.
>
OK, so even if different users get the same "total lineup" hash (which
means some users will re-download days where none of the programming on
the channels in their lineup has changed), TMS /still/ has to generate a
hash. Generating a hash takes processing resources (of which TMS is
limited). When you think about how many lineups they would have to hash
after every single listings change, it's not insignificant--even if they
only do it for "total lineups" instead of user-specific lineups. If NBC
changes a single program, how many thousands of "total lineups" are
affected?
In truth, it would make more sense to do a revision number. But, I
suspect doing so would significantly increase the bandwidth we
(non-paying) DataDirect users would consume. And, I would guess a
significant portion of the additional bandwidth would be for downloads
of lineups that aren't affected by the schedule change. (I have 13
channels in my custom lineup. There are 903 channels in my total
lineup. How many times do you think a schedule change on 1 of 903
channels would happen to be one of my 13 channels?)
> It's the same San Francisco Bay
>Area OTA HD channel-set that hundreds or thousand of people in this
>area share. The burden of removing the channels I don't want could
>either be shifted to the client (meaning slightly larger downloads),
>
>
TMS made a conscious decision to give custom lineups at the expense of
processing resources to save bandwidth. I doubt they would be willing
to go back on this decision.
>or the download-method could be changed to be per-channel (meaning
>many small downloads, but only for those channels that changed).
>There are plenty of ways to solve this problem, I think, but the
>question is whether zap2it wants to invest any time in this.
>My guess is no.
>
And this is not the place to discuss it (because no one here can do
anything about it). This is: http://bb.labs.zap2it.com/ .
<soapbox>
It's amazing what people will complain about even when they get a
commercial-quality service (the same one TiVo users pay to get) for
free. Before long people will be complaining that they downloaded their
listings at 7:45 and TMS received (from the network) at 8:00 a schedule
change for tomorrow, so they never got it, and then they'll (again) say
that TMS should be providing an RSS feed or ...
</soapbox>
Remember, folks, it's just TV. And, to put things in perspective, think
about how much nicer we DataDirect users have it than users in the rest
of the world.
Mike
More information about the mythtv-users
mailing list