[XCON] Draft Minutes from XCON WG, IETF58, Wed 12nov2003 15:30-17:30

[Paul Kyzivat <[e-mail]>]

Reviewed charter milestones. Noted that this is first meeting of the WG 
and we're already behind.

[Ed: There was an initial question about something, but I couldn't 
understand what was said or the answer. I think it was by Roy Radhika - 
something about plans for extensions to conf control.]

Discussed draft-even-xcon-conference-scenarios-00. There were questions 
about adequacy of scenarios, especially regarding text media.
* Cullen Jennings agreed to contribute some text on that subject.
* On humm, agreed to accept the document as WG item.

Joerg Ott discussed draft-koskelainen-xcon-floor-control-req-00. There 
was question by Jonathan Rosenberg about relationship to focus - not 
covered by document. Roni Even and Rohan Mahy also asked about 
positioning re the conferencing framework. (There seemed to be sentiment 
for better positioning of this doc, but no humm on it.) Juhe Garg raised 
concern with the number of protocols that conference participants must 
implement. Answer was that this doc is just requirements, not 
specification of a protocol. Nermeen Ismail asked whether participants 
are notified who has the floor. Answer was there is a requirement for 
this. Cullen Jennings asked for way for requestor to know what status is 
of request for floor.

Marcus Brunner discussed draft-brunner-xcon-fc-issues-00. This draft 
asserts floor control mechanism must be independent of policy; policy is 
instead part of the conference application. Question by chair if this 
could be incorporated into the floor control requirements document. 
Joerg and author had discussion of fine points of this. Decided this 
should be taken to the mailing list. Brian Rosen expressed opinion that 
the requirements were verging on too much for initial work.
* On humm, agreed to accept the combination of these two documents as WG 
doc.

Hisham Khartabil presented draft-koskelainen-xcon-cpcp-reqs-01. Juhe 
Garg requested addition of requirement to control when a conference ends.
*Hummed to adopt this as WG item.

Hisham also discussed draft-koskelainen-xcon-xcap-cpcp-usage-01: asked 
for opinions on use of XCAP for this. The people present were not 
prepared to commit on this.

Orit Levin presented draft-levin-xcon-cpcp-00 (apparently an alternative 
to draft-koskelainen-xcon-xcap-cpcp-usage-01.) She had some issues with 
XCAP - expressed opinion it would unduly influence schema - preferred 
focusing on xml schema independent of transport used to manage it - open 
to http, sip, soap. Discussion continued on both documents. Cullen 
Jennings thought things don't hang together very well yet - need more 
work. Alan Johnston complained that he hadn't seen anything on list 
about it - asserted not ready to decide on one approach vs the other - 
need to take this to the list. Requested people to read and comment on 
the requirements.
* Cullen Jennings, Eric Burger, and others agreed to do so.

Rohan discussed draft-mahy-xcon-media-policy-control-00. It raises issue 
of what level of abstraction does the wg want to work at. There was very 
lively discussion of this, only slightly captured here. Got to question 
of whether the application deals with application specific roles, and 
they are the subjects in intereaction with media policy, or whether the 
application must itself map its notion of role onto some 
non-application-specific notation of media streams, etc. Cullen Jennings 
thought roles of participants can change - wasn't sure if everybody on 
same page. Doesn't want it to be hard to figure out roles. He also also 
doesn't know how an "application" relates to the model, but apparently 
it is important. Nermeen Ismail reiterated desire to make policy engine 
unaware of significance of particular roles. Eric Burger wanted 
something fairly high level - gave analogy - assembly language is more 
flexible than Java, but wojuld prefer to use Java. Rohan and Eric 
debated on connections between media and conference policy. Rohan thinks 
there are cases where changes to role of a participant in conf policy 
also affect media policy. Eric says "not necessarily" - it is up to 
"application" to decide if both should be changed. Alan raised issue 
that changes to media policy must be able to be reflected back to users 
- need some way to reverse engineer the changes to something user 
meaningful. Thought that the conceptual approach was better than the 
logical (Rohan) approach for this. Juhe Garg wanted to remove term 
"roles" - focus only on "permissions". Eric really didn't like this. 
Rohan felt use cases are needed. Nermeen wanted better requirements. 
Rohan then showed a strawhorse def of a subconference - asserting it is 
a conversation.

Brian Rosen then showed a couple of slides 
(jennings-xconMediaPolicyRan.pdf) of rebutal to Rohan's, asking what is 
the problem we are trying to solve. Asserted 
draft-mahy-xcon-media-policy-control-00 is too complicated, and doesn't 
address the problem directly. Its a language to describe media flow 
graphs, not one to control them. Not facilitating interoperable 
implementations. Brian proposed templates instead. Rohan was concerned 
with how to map templates for media control onto conference control. 
Brian gave an example of a mix-minus mixer. Said #input audio streams, 
#output audio streams (max) are bound at conf creation. This raised a 
lot of concerns from audience. People felt this was almost what Rohan 
was proposing. Rohan asserted that any proposal must address how it 
interacts with conf policy. And Cullen wanted to be sure it covers all 
the requirements. Jonathan didn't think it was necesssary to solve all 
problems as long as there is extensibility.

Media policy requirements draft by Roni Even mentioned: hasn't changed 
since last meeting.

That's it.

	Paul Kyzivat


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