Re: Where are we?

From: Jack Lang (fw23@dial.pipex.com)
Date: Sun, Oct 10 1999


Message-ID: <005e01bf130e$515ecb20$0a8db40a@ntl.com>
From: "Jack Lang" <fw23@dial.pipex.com>
To: "Dan Zigmond" <djz@corp.webtv.net>, "'Scott J. Anderson'" <sjanderson@newshour.org>, <www-tv@w3c.org>, "Harald Tveit Alvestrand" <Harald@Alvestrand.no>
Date: Sun, 10 Oct 1999 11:57:57 +0100
Subject: Re: Where are we?

Harald:

Unfortunately the slot has been delayed by ten minutes, as the earlier news
program was extended.

Although I support the notion of the ability to add future extensions to the
naming scheme, I  think it's too hard to specify them at the moment. In
particular, time is very complex in TV land, and there is considerable
ongoing discussion on the subject in the trigger debates.  Even for one
channel, one physical location and local time (as in your example) there can
be significant delays depending on the transmitter and transmission method.
The same channel might be available from cable, satellite, and several
terrestrial transmitters, both analogue and digital, and all with different
timings compared to origination.
End to end reliable time, such as NPT, doesn't usually make through the
transmission chain, nor do indications of schedule slippages.

At the least there is

Absolute time (e.g. UTC)
Local time
Original (EPG) schedule time
Actual schedule time
Current transmission information (e.g SI tables)
NPT

There are also various internal times, depending on the transmission method,
such as Teletext time, and time bases within the DVB and ATSC encodings.
There are also complications associated with VOD and NVOD channels, and with
local DVCR functions like pause.

In view of this I'm not sure why its useful to specify a time-slot
independent of the program that might occupy it. I can understand saying
things like "I want to record Baywatch on BBC1", or "Pop up this page 5 mins
38 seconds after the start of Baywatch Episode 13 on BBC1", but tying it to
a specific timeslot seems dangerous.

Jack Lang


> 2) As a TV watcher, when I've identified a channel, the next level of
> identification is commonly a timeslot, as is done by the ShowView
> bizarre-digit scheme here in Norway, for instance.
> I would regard content as being an orthogonal identifier, but timeslot
> seems intrinsically channel-bound.
>
> Not that it's simple - see the CALSCH calendar specifications for just how
> complex "every thursday at 9 AM, here in Trondheim" can be to specify
> exactly - but it's a logical extension to a channel/program stream
> identification scheme. (did we ever get down to writing up a glossary for
> this stuff??)
>
> Making sure it's possible to extend the scheme in that direction may be
the
> only thing we should do now.
>
>                          Harald
> --
> Harald Tveit Alvestrand, Maxware, Norway
> Harald.Alvestrand@maxware.no