[opendtv] Re: IP-Based TV Will Revolutionize Entertainment

  • From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
  • To: "OpenDTV (E-mail)" <opendtv@xxxxxxxxxxxxx>
  • Date: Sun, 1 May 2005 18:23:30 -0400

Venki Iyer wrote:

> but I read this as a non-tech (well, OK, non-IP, non-XML,
> non-RSS, take your pick) editor/writer trying to describe
> what can and is being done with RSS (which uses XML) and
> IP-based protocols (like BT, for example).

Except that the part "and IP-based protocols" is exactly what
we are trying to identify. To say "IP-based protocols" is
*not* to say that mentioning IP is enough.

IP is a routing protocol, layer 3, described in RFC 791. It
doesn't begin to specify what an IPTV (or any other DTV)
needs to interoperate with client hosts. Kon may be correct
that all IPTV schemes use XML for layers 5-7, I have no way
of knowing. I've never seen a comprehensive survey of IPTV
systems deployed so far. (But I would be amazed if they were

The same applies to ATSC and DVB-T. It's hardly enough to
limit the definition to MPEG-2 TS. For exactly the same
reasons. How does the client interpret the bit stream?

Even the new media darling, Bit Torrent, doesn't get away
with any of this. BT is essentially a layer 5 scheme, session
layer. It works on top of TCP, and it orchestrates a
choreography of TCP downloads from a number of different
servers. The question remains, after the download is
reconsituted into the appropriate file, what have you
downloaded and how is it read? Is it a big MS Word file? Is
it Acrobat? Is with WM10? Heck, is it ATSC over MPEG-2 TS
(why not?).

You need this info for clients to decode the stream they're
receiving. And you are also making assumptions of how
powerful the client hardware is, whether it's cabable of
coping with this bit stream.

> Maybe this makes Kon's point, that IBM and Harris are not
> exactly the shining examples of the leading edge here,

Kon focused on XML as the choice. I don't think that was
the thrust of Loring Wirbel's column. As I indicated, he
was saying that only Harris and Microsoft seemed to grasp
that merely specifying "IP" is simply not enough. And I
tried to make the point more clearly by tying this back to
the discussion of what it takes to define a DTT system. The
concepts are the same, whether it's IP or other digital
protocol at layers 2 and 3.

The ATSC and the DVB have devised a very efficient way to
fully define DTT streams. Furthermore, with A/90 and the
DVB equivalent (whatever that is), they have opened the
doors to almost limitless options for upgrading the
standard, subject only to the one-way broadcast environment
DTT operates in. In principle, even XML and some one-way
version of Bit Torrent could be transmitted over ATSC and
DVB-T. (Imagine for example a tuner being told to tune
simultaneously to x different stations, download the
separate streams, store in a PVR, reconstitute the original
file, then interpret the file.)

So, my objections continue to be (1) that a vague mention
of "IP" is not enough to describe a TV systems that works,
(2) that the same applies to MPEG-2 TS for a TV system,
and (3) the claim that ATSC is "not extensible" is way off
the mark.

You can UNSUBSCRIBE from the OpenDTV list in two ways:

- Using the UNSUBSCRIBE command in your user configuration settings at 

- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: