Re: [Simple] MSRP chat: draft-niemi-simple-chat-04.txt

[Miguel Garcia <Miguel.An.Garcia at nokia.com>]

Paul:

Inline comments:

Paul Kyzivat wrote:
> This is a great improvement. My remaining comments are primarily about 
> nicknames:
>
> - as you already comment, I think nicknames belong as a feature of 
> conferences in general, not MSRP mixer sessions. In principle, I think 
> the assignment of nicknames should be handled via the conference control 
> protocol. (Whose accronym I currently forget.) As a result I am troubled 
> by using MSRP as a protocol for doing this. (Should we also extend RTP 
> to support assigning of nicknames?) The implications of doing this then 
> cascade down to having new SDP to negotiate whether the MSRP peer is 
> capable of assigning nicknames.

The current idea is that nicknames are especially useful when you use 
them for providing privacy, but still be able to send "a communication" 
to the other party via the conference server. This is easy to fulfill 
with any Instant Messaging protocol, because if I the conference server 
anonymizes me, I can still send you an IM and you can reply back to me 
through my nickname, but you don't know my real SIP URI.

I don't really know what is the use case for RTP, because if I send you 
audio or video you will identify me immediately, in spite the conference 
server is anonymizing me.

So that has been the main driver for providing the function in MSRP. The 
other thing is that, existing systems (and IRC is an example of it) 
provide similar functionality.

Yet another thing... I am not sure what will be the reaction of the SIP 
WG if we propose the standardization of the NICKNAME method...


> - (minor editorial issue) I think you intend to permit other mechanisms 
> to deal with nickname retention. There is some language to that effect. 
> But there is also language that brings this into doubt:
>
>    In other words, a nickname is simply a username that is scoped for a
>    particular chat room.  Such nicknames are allocated on a first-come
>    first-served policy, meaning they can also be "stolen".  It is out of
>    the scope of this specification to define nickname retention schemes,
>    or nickaming services as discussed above.
>
> This seems too strong about what a nickname *is*. I think it would be 
> good to say that the server *may* issue nicknames scoped this way. But 
> the user should be able to suggest any nickname that it believes the 
> server may allow it to use, and the server may accept it. This would 
> give much more latitude for future work.

OK, sounds a nice suggestion.

> - The lifetime of the nickname is said to be the lifetime of the 
> session. But it isn't clear if that is the MSRP session or the SIP 
> session. (IMO it ought to be the SIP session.) One might retain the sip 
> session and change the MSRP session. This really highlights whether the 
> nickname is chat specific.

This is a good point since, I always got in mind that the SIP session 
has an MSRP session... but I see the difference, so obviously that has 
to be clarified. And I see the point that this should be tied to the SIP 
session...


Thanks,

     Miguel


> That's all.
>
>     Paul
>
> Miguel Garcia wrote:
> > We have just submitted a new version of the MSRP CHAT draft. This has 
> > changed substantially from earlier versions, so I wouldn't dare to 
> > summarize the changes.
> >
> > Until the document is published, you can fetch a copy from:
> > http://people.nokia.net/~miguel/drafts/draft-niemi-simple-chat-04.txt
> > http://people.nokia.net/~miguel/drafts/draft-niemi-simple-chat-04.html
> >
> > Comments are welcome.
> >
> > BR,
> >
> >
> >         Miguel

--
Miguel A. Garcia           tel:+358-50-4804586
sip:miguel.an.garcia at openlaboratory.net
Nokia Research Center      Helsinki, Finland


_______________________________________________
Simple mailing list
Simple at ietf.org
https://www1.ietf.org/mailman/listinfo/simple