Hi, I agree with John. The 'label' RFC is essentially syntax. It seems to me this source/sink issue has some things in common with the work in CLUE (formerly maitai) in the sense that it too has to face the issue of linking devices with streams... It may be that they need a common protocol mechanism... Regards, Peter Musgrave On 2011-01-10, at 2:20 AM, Elwell, John wrote:
-----Original Message----- From: Harald Alvestrand [mailto:harald@alvestrand.no] Sent: 08 January 2011 14:03 To: Elwell, John Cc: 'dispatch@ietf.org'; rtc-web@alvestrand.no Subject: Re: [RTW] Charter proposal: The activity hitherto known as "RTC-WEB at IETF"
On 01/07/11 10:38, Elwell, John wrote:
Harald,
With one exception, this draft charter does not mention signalling (SIP/SDP), and I assume it is intended to be out of scope, although that is not explicitly stated.
The one exception is "8) RFC 4574 based Label support for identifying streams purpose". This is an SDP attribute. How would this be used if SDP is indeed out of scope? If SDP is in scope, we certainly need more text describing its relevance. There isn't much text in 4574 - the main point of the RFC seems to be that it's possible to create labels for streams before creating the stream. Those labels could be independent of the signalling mechanism.
The trend I see now is that most APIs and protocols people define (for instance Jabber/XMPP, JSON-based descriptions) have some kind of mapping to SDP, so SDP-as-a-format may not be so important (I heartily dislike multiple things about the specific format), but SDP-as-a-semantic becomes more important over time, and can't be ignored. [JRE] RFC 4574 is little more than an SDP syntax (for conveying an application-specific label). As such, if we don't use the SDP syntax, I don't see how RFC 4574 can apply at all. The problem we need to solve, if I understand correctly, is for the application to instruct the browser what source and sink to use for a given medium. For audio, this means telling the browser which microphone and which speaker to use. For video, this means telling the browser which camera to use and where on the display to render received video. I don't see the relevance of RFC 4574 for achieving this.
John
Harald
_______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch