Re: [clue] Design team meeting on Tuesday

[Robert Hansen <[e-mail]>]

Here are the notes from the meeting:

IETF CLUE design team meeting, 19/06/2012

Participants:

Paul Kyzivat
Brian Baldino
John Leslie
Robert Hansen
Roni Even
Allyn Romanow
Andy Pepperell
Mary Barnes

Allyn raised the question of whether it would be possible to change the 
Webex options to generate a toll-free number to dial in with. Paul said 
he would look into it. Roni pointed out it was possible to join the 
conference via the PC.

It was decided to discuss draft-even-clue-swithed-use-cases-00 - as some 
people had not read it Roni shared the document.

Rob asked if when the document says that the document says the framework 
defines "mixed" it meant "switched" - Roni confirmed it did.

Roni went through the example in the introduction, first with the 
example of an endpoint offering the four capture scene entries, then an 
MCU doing much the same. He pointed out that in the MCU case, the first 
capture scene entries might be switched. Paul asked if, in this case, 
the captures would still have spatial information. Roni said it would be 
up to the MCU, but that generally they would, but not with 'real' 
coordinates.

Paul asked whether a device such as the one given in the example was 
obligated to give a sensible two-screen option. Roni said that no, and 
that at least in the initial case the provider had no concept of how 
many screens the consumer had. If the provider doesn't offer anything 
optimal the consumer will have to do the best it can from the captures 
available.

Roni suggested that based on what is currently in the framework, there 
is some limited ability for a provider to adjust its advertisment based 
on the other side's advertisment. Rob suggested that there isn't 
necessarily symmetry between what a device can send and what it can 
receive. Roni agreed that this wasn't necessarily an optimal approach, 
but was what could be done using the current framework.

Paul suggested that when the number of screens was small it would be 
relatively easy for a provider to advertise all sensible combination of 
screens, but as the number of screens increases this becomes less 
sensible as the problem becomes combinatorial. Rob pointed out that as 
the number of screens increases the potential ways they can be laid out 
all increases (with three or fewer screens

Allyn suggested at this point that with so little time remaining in the 
call we should move further on in the document. There was some 
discussion of what should be discussed next.

Roni said he thought that the current framework was sufficient to work 
for telepresence systems with one to three screens laid out left to 
right, with the number of cameras matching the number of screens. There 
are then other scenarios (number of monitors not matching number of 
screens, non-linear layout of screens, endpoints doing local 
composition, etc) which are not currently solvable with the current 
framework, and there is a trade-off between allowing more scenarios and 
the time required to complete the process. Allyn expressed the fact that 
this wasn't necessarily an all-or-nothing thing, and that some of the 
functionality could be added to the framework while others were not.

There was some discussion about the format of the use cases; Allyn and 
Andy felt some of the use-cases in the document were more mechanisms for 
solving the problem than they were pure use-cases. Roni explained that 
when writing the user cases some of them were linked and could 
potentially be joined together. Allyn said she would send some notes on 
the user-cases to the list.

Roni asked for the interim meeting minutes - Mary stated that they had 
for the most part been published, and that she would send a link to the 
list.

Paul asked if there were any other user-cases, as whether there were any 
cases involving the N+M mix of main and presentation video. Roni said 
that this was a general case of one case that had already been 
documented with the simple case of a three screen system taking two main 
video and one presentation video. Paul was interested in what the 
behaviour should be when a consumer has the option of subscribing to 
main and presentation and doesn't have the resources to receive both, 
and whether it was generally true that presentation should be 
priorities. Mary asked Paul to send a note to the list on the subject.

On 18/06/2012 19:39, Mary Barnes wrote:
> Hi all,
>
> I propose that we discuss Roni's use cases tomorrow:
> http://www.ietf.org/mail-archive/web/clue/current/msg01597.html
>
> As a reminder here are the Webex details:
> http://trac.tools.ietf.org/wg/clue/trac/wiki/Design-Team
>
> Thanks,
> Mary.
>
>
> _______________________________________________
> clue mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/clue