Reminder to self: "reply list" is not a wise move when discussing on 2 lists simultaneously... -------- Original Message -------- Subject: Re: [dispatch] Charter proposal: The activity hitherto known as "RTC-WEB at IETF" Date: Wed, 12 Jan 2011 14:43:11 +0100 From: Harald Alvestrand <harald@alvestrand.no> To: dispatch@ietf.org On 01/12/11 13:49, Alex Eleftheriadis wrote:

I think this is an interesting and important activity, but I want to point out a few points of concern with the current proposal.

Although the charter indicates that there will be a "selection", it fails to identify the criteria for this selection process. It is easy to see that this can lead to a shouting match. The only indication of a criterion for the codecs is:

...

* interoperate with compatible voice and video systems that are not web based

which, btw, you already re-interpreted to mean "brower-to-browser" in a later email. Aside: I'm worried about this bullet in the charter; I'm not convinced it's either clear enough or saying the right thing. I know we have to do browser-to-browser; unless we can define "compatible voice and video systems" in a way that people have a common understanding of, we might be better off dropping the line from the charter or moving it to the section "work that we're not doing but might look at after a recharter". I would therefore recommend that, if a single baseline codec is to be selected, the criteria are stated in the charter. If the criteria are to be defined later by the group, then that should be stated in the charter, with at least some indication of what the group wants to accomplish here. I don't think we can get the criteria in the charter. In fact, I wonder if we have to add a criteria document to the deliverables. As an example, a statement was made yesterday that Theora's delay characteristics make it "unsuitable for interactive use" - while I have little reason to disbelieve that statement, a discussion of exactly how many ms of the ~200 ms latency budget of a typical interactive video call we can allow the codec to consume can easily consume many messages and many days/weeks of time.

I think it's important that the charter states that there should *be* a mandatory-to-implement, and that we're envisioning evaluating existing codecs rather than inventing new ones. Suggestions for appropriate text? _______________________________________________ dispatch mailing list dispatch@ietf.org https://www.ietf.org/mailman/listinfo/dispatch