RE: [MMUSIC] I-D ACTION:draft-ietf-mmusic-sdp-media-content-04.txt

[Gunnar Hellström <[e-mail]>]

Cullen,
Thanks for giving the media-content the required decision to move on.

This gives us a reason to review the requirements behind the proposed "txp"
parameter for text turntaking.
It can in fact be a combined audio/text turn-taking procedure. Real-time
text and voice takes turns during a call in many cases on the PSTN side. And
then, during text periods, turn-taking is often used for text in the two
directions.

That leads me to think that we should look at session level attributes.
Such as
Contact: duplex="half" or a suitable more precise indication in SIP.

It should be remembered that the media flow itself should be left to be full
duplex on the IP side.

More on this soon under another subject line.

Gunnar

----------------------------------------------------------------------------
-
Gunnar Hellström, Omnitor
[e-mail]
Mob: [phone]
Phone: [phone]
www.omnitor.se


-----Original Message-----
From: Cullen Jennings [mailto:[e-mail]]
Sent: Friday, August 25, 2006 6:00 PM
To: Arnoud van Wijk
Cc: 'MMusic'; 'Gunnar Hellström' ; 'Gonzalo Camarillo'; 'Colin Perkins'
Subject: Re: [MMUSIC] I-D
ACTION:draft-ietf-mmusic-sdp-media-content-04.txt



Ok, folks, with my AD hat on here, here is what I would appreciate it
if you would do:

Remove these two attributes from this draft - either we have come to
consensus they should not be here or we have at least agreed that we
don't have consensus on it.

Update the description of what are appropriate future extensions of
this to include text along the lines of Gonzalo's

  "Content tokens should consist of a meta description of the
contents of a media stream. That is, whether the stream contains
slides, or is the main stream, etc. Content tokens should not
describe things like what to do in order to handle a stream (e.g.,
use user-level floor control)."

Resubmit this draft soon and I will start IETF LC on it. The WG can
provide any comments that might typically be WGLC comments in IETF LC.

Go to work on another draft that solves the txp, user-floor problem -
I will do what I can to help expedite that draft. One thing  that I
think will help is if the draft is clear about what happens  when a
device that receives this signally and how that changes the receiving
devices behavior.

Thanks, Cullen


On Aug 25, 2006, at 1:43 AM, Arnoud van Wijk wrote:

> Very well.
> That is clear then.
> We shall focus on the separate SDP attribute.
>
> Arnoud
>
> Arnoud A. T. van Wijk
> AnnieS New Technology
> www.AnnieS.nl
> Mobile textphone (SIP): [e-mail]
>
>> -----Original Message-----
>> From: Gonzalo Camarillo [mailto:[e-mail]]
>> Sent: vrijdag 25 augustus 2006 10:02
>> To: Arnoud van Wijk
>> Cc: 'Colin Perkins'; 'Gunnar Hellström'; 'Cullen Jennings'; 'MMusic'
>> Subject: Re: [MMUSIC] I-D ACTION:draft-ietf-mmusic-sdp-media-content-
>> 04.txt
>>
>> Hi,
>>
>> my take on this would be to remove the 'user-floor' and the 'txt'
>> tokens
>> from this draft and add a clearer explanation of what the content
>> tokens
>> (current and future) should describe.
>>
>> Content tokens should consist of a meta description of the
>> contents of a
>> media stream. That is, whether the stream contains slides, or is the
>> main stream, etc. Content tokens should not describe things like
>> what to
>> do in order to handle a stream (e.g., use user-level floor control).
>> That type of stuff would be better described in a separate SDP
>> attribute, as suggested by Colin at some point.
>>
>> That separate SDP attribute would resolve the use cases Arnoud and
>> Gunnar have in mind, which are very interesting as well.
>>
>> Thanks,
>>
>> Gonzalo
>>
>>
>> Arnoud van Wijk wrote:
>>> Hi Colin,
>>> I agree with Gunnar and disagree with you.
>>> The txp content label is no different the the other content
>>> labels IMHO.
>>> And again, it is NOT about a PSTN textphone, but about turntaking be
>> used in
>>> the call since it media stream contains real-time text to be used in
>> turn
>>> taking method.
>>> It is not about a special phone, or limited device that is elsewise
>> unable
>>> to be connected.
>>> But I am still waiting on Gonzalo's view on this.
>>>
>>> Greetz
>>>
>>> Arnoud
>>>
>>> Arnoud A. T. van Wijk
>>> AnnieS New Technology
>>> www.AnnieS.nl
>>> Mobile textphone (SIP): [e-mail]
>>>
>>>> -----Original Message-----
>>>> From: Colin Perkins [mailto:[e-mail]]
>>>> Sent: woensdag 23 augustus 2006 10:15
>>>> To: Gunnar Hellström
>>>> Cc: Cullen Jennings; MMusic; Gonzalo. Camarillo@ericsson. com
>>>> Subject: Re: [MMUSIC] I-D ACTION:draft-ietf-mmusic-sdp-media-
>>>> content-
>>>> 04.txt
>>>>
>>>> Gunnar,
>>>>
>>>> On 23 Aug 2006, at 08:28, Gunnar Hellström wrote:
>>>>> It is allowed to have more than one content attribute.
>>>> True, but irrelevant.
>>>>
>>>>> There is no difference in characteristics of the txp attribute
>>>>> compared to the other accepted attributes.
>>>>> slides come from a slide showing application that has a specific
>>>>> expected behaviour of showing one slide at a time.
>>>>> Speaker has its source in a camera, a keyboard and a microphone
>>>>> that likely requires some polite method to be allowed to control.
>>>>> All have sources in devices and have a mixed meaning of device
>>>>> indications and preparation for human behaviour when using the
>>>>> contents.
>>>>> There is no reason to separate out txp in this collection of
>>>>> attributes.
>>>> I disagree, for the reasons I have previously explained.
>>>>
>>>>> With the new slimmed description it is of the same size and level
>>>>> as the others.
>>>> The "size and level" of the description was never the problem.
>>>> Instead, there is concern that its semantics don't fit with the
>>>> content label.
>>>>
>>>>> As I remember it, the last call just asked for an improved
>>>>> description, not questioning the idea to include the attribute.
>>>> I am explicitly questioning the inclusion of the txp label in the
>>>> media content draft. I believe it should be removed to a separate
>>>> draft, defining a new SDP attribute. We should also consider if the
>>>> user-floor label should be removed too.
>>>>
>>>> A simple rewording of the description of the txp label, of the type
>>>> you are proposing, without considering the semantics, will not
>>>> resolve this issue.
>>>>
>>>> Colin
>>>> _______________________________________________
>>>> mmusic mailing list
>>>> [e-mail]
>>>> https://www1.ietf.org/mailman/listinfo/mmusic
>>>
>>>

_______________________________________________
mmusic mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/mmusic




_______________________________________________
mmusic mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/mmusic